From steve.lime at dnr.state.mn.us  Fri Oct  1 11:38:42 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 01 Oct 1999 13:38:42 -0500
Subject: mapserver examples/tutorial
Message-ID: <s7f4b96e.076@smtp.dnr.state.mn.us>

Hi: As part of a workshop I'm doing I setup a simple progression of
mapserver interfaces online. They are simple, but they show enough
stuff to be useful to some folks, I hope. Each example is accompanied
by corresponding map files and all templates. You'll need Netscape
to access the templates as MSIE likes to parse everything that is 
text regardless of MIME type. Here's the URL to the first example:

http://www.dnr.state.mn.us/mis/gis/gislis99/example1.html

There are 5 in the progression. Since these are for use in a talk there's
little text on them so here's what to look for in the progression.

example 1: - single layer called from an <img> tag, no template so view source

example 2: - multiple layers called from an <img> tag, no template so view source

example 3: - basic interface, adds pan and zoom
                  - sticky zoom
                  - adds legend, scalebar and reference map

example 4: - more advanced interface, adds sticky layer selection
                  - mapfile introduces scale dependent lake layers

...big jump from to 5...

example 5: - adds java applet for box area of intrest selection
                  - adds query (mode=nquery), user can select 1 or many (with box)
                  - auto legend replaced with cleaner version (you'll see what I mean)
                  - javascript interface logic
                  - nice example of using javascript to handle query output, does one thing
                    if only one feature is found, and another if more than one is found

Server is running 3.2.21. All data is at http://www.dnr.state.mn.us/mis/gis/gislis99/data
and graphics are at http://www.dnr.state.mn.us/mis/gis/gislis99/graphics in case any
one wants to download stuff to play with themselves. Applet is available on request.

Hope it helps. 

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us



From steve.lime at dnr.state.mn.us  Fri Oct  1 13:40:41 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 01 Oct 1999 15:40:41 -0500
Subject: 3.2.21 Available.
Message-ID: <s7f4d60c.000@smtp.dnr.state.mn.us>


Version 3.2.21 is available at the MapServer website. Not a major
release at all. Has 3 fixes:

  - Fixes bug with error traps. If parameters are wrong in map file then
    the standard ugly old error format is used. Only errors that occur 
    after initial map file parsing are catchable.

  - Fixed bug that caused one-to-many join processor to kick in cases
    where an attribute table contained an item name that's a substring of
    the join name. Thanks Ken.

  - Added DD (decimal degrees) as a valid unit type. You can now use
    scale related parameters MINSCALE, MAXSCALE, SYMBOLSCALE
    and scalebars with lat/lon data. Remember though, the computed
    scale is only accurate at the equator and should be used internally
    (i.e. within a map file) only. DD is NOT a valid scalebar unit.

Steve



From cshorter at optusnet.com.au  Fri Oct  1 14:10:51 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Sat, 02 Oct 1999 07:10:51 +1000
Subject: mapserver examples/tutorial
References: <s7f4b96e.076@smtp.dnr.state.mn.us>
Message-ID: <37F5235B.AA1D0CF9@optusnet.com.au>



Stephen Lime wrote:
> 
> Hi: As part of a workshop I'm doing I setup a simple progression of
> mapserver interfaces online. They are simple, but they show enough
> stuff to be useful to some folks, I hope. Each example is accompanied
> by corresponding map files and all templates. You'll need Netscape
> to access the templates as MSIE likes to parse everything that is
> text regardless of MIME type. Here's the URL to the first example:
> 
> http://www.dnr.state.mn.us/mis/gis/gislis99/example1.html

Great.  How about you link it into the mapserver doco pages:
http://mapserver.gis.umn.edu/doc.html

I'm having problems running example 5.  I got "Applet mapimage can't start:
ERROR".

I'm running Netscape Communicator version 4.6.



From rchavez at perseus.tufts.edu  Tue Oct  5 12:34:15 1999
From: rchavez at perseus.tufts.edu (Robert Chavez)
Date: Tue, 05 Oct 1999 15:34:15 -0400
Subject: mapserver and database connectivity.
References: <s7f4d60c.000@smtp.dnr.state.mn.us>
Message-ID: <37FA52B7.5343F019@medusa.perseus.tufts.edu>


Greetings,

I am wondering if it is possible to send information to mapserver from a
database (PostGresql).

 I'm using the Pg perl module to build a web interface to my PostGres
database so users can query and display the data set.  The next logical step
would be to send the results to Mapserver somehow and plot them on my
basemap -- or perhaps create a tranparent gif of the coordinates that could
be overlayed on the base map.  Has anyone had any experience  with this sort
of functionality?

Thanks,
Rob Chavez




From bowenj at gaiaenv.com  Tue Oct  5 13:11:16 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Tue, 05 Oct 1999 14:11:16 -0600
Subject: mapserver and database connectivity.
References: <s7f4d60c.000@smtp.dnr.state.mn.us> <37FA52B7.5343F019@medusa.perseus.tufts.edu>
Message-ID: <37FA5B64.F510D0C4@gaiaenv.com>

Hi all,

Second this question.  We are looking at building similar functionality,
but with PHP.  When we discover any info. re. this, we will post it. 
Perhaps Steve has some insight?

Cheers,
Jim

Robert Chavez wrote:
> 
> Greetings,
> 
> I am wondering if it is possible to send information to mapserver from a
> database (PostGresql).
> 
>  I'm using the Pg perl module to build a web interface to my PostGres
> database so users can query and display the data set.  The next logical step
> would be to send the results to Mapserver somehow and plot them on my
> basemap -- or perhaps create a tranparent gif of the coordinates that could
> be overlayed on the base map.  Has anyone had any experience  with this sort
> of functionality?
> 
> Thanks,
> Rob Chavez

-- 
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA


From ncarr at guideguide.com  Wed Oct  6 03:54:47 1999
From: ncarr at guideguide.com (Nathan Carr)
Date: Wed, 06 Oct 1999 11:54:47 +0100
Subject: mapserver and database connectivity.
References: <s7f4d60c.000@smtp.dnr.state.mn.us> <37FA52B7.5343F019@medusa.perseus.tufts.edu> <37FA5B64.F510D0C4@gaiaenv.com>
Message-ID: <37FB2A77.BB41A552@guideguide.com>

	I too am interested in this functionality. Have been looking at
wrapping
mapserver with SWIG, so I could use perl DBI to access database tables.
This seems a bit
confusing as there is quite a bit of libraries and structures to
include.
 
Alternatively would it be possible to extend the JOIN functionality for
joining DBF
files to join MySQL/PostGres type tables? 

Cheerio, nathan

> 
> Hi all,
> 
> Second this question.  We are looking at building similar functionality,
> but with PHP.  When we discover any info. re. this, we will post it.
> Perhaps Steve has some insight?
> 
> Cheers,
> Jim
> 
> Robert Chavez wrote:
> >
> > Greetings,
> >
> > I am wondering if it is possible to send information to mapserver from a
> > database (PostGresql).
> >
> >  I'm using the Pg perl module to build a web interface to my PostGres
> > database so users can query and display the data set.  The next logical step
> > would be to send the results to Mapserver somehow and plot them on my
> > basemap -- or perhaps create a tranparent gif of the coordinates that could
> > be overlayed on the base map.  Has anyone had any experience  with this sort
> > of functionality?
> >
> > Thanks,
> > Rob Chavez
> 
> --
> 
-- 
Nathan Carr  ncarr at guideguide.com


From teb at mallit.fr.umn.edu  Wed Oct  6 08:14:15 1999
From: teb at mallit.fr.umn.edu (Thomas E. Burk)
Date: Wed, 6 Oct 1999 10:14:15 -0500 (CDT)
Subject: fwd: Map Server Test Page up
Message-ID: <199910061514.KAA27471@mallit.fr.umn.edu>

Forwarded with permission of the author.

Tom

------------- Begin Forwarded Message -------------


From pgallen at randomlogic.com  Mon Oct  4 22:54:32 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 04 Oct 1999 22:54:32 -0700
Subject: Map Server Test Page up
Message-ID: <mailman.150.1728305905.23453.mapserver-users@lists.osgeo.org>

I have fixed my firewall and my demo pages should now be available to all (I've
had a couple people tell me they were successful at browsing the pages). Once
again the URL is:

24.5.14.144
http://cx570538-a.elcjn1.sdca.home.com

If you're wondering why it's a strange URL or IP: I'm on a cable modem (1MB/sec.
connection) using IP masquerading and forwarding on Red Hat Linux 6.0. Even
though I have 3 legitimate web sites, none of them are located on my home
machine and doing it this way gave me a free and easy method for testing without
having to ftp changes all the time.

Question:

Is there a resource for up to date, low cost (or free) map databases? One of the
issues my company has is the availability of current map data for the United
States. MapQuest provides it for $20,000/year (and an additional $50,000 for the
map server!). If I can find a low cost solution, that'd be great.

PGA
------------- End Forwarded Message -------------



From TBanister at SpaceImaging.com  Wed Oct  6 09:38:59 1999
From: TBanister at SpaceImaging.com (Banister Trevor)
Date: Wed, 6 Oct 1999 10:38:59 -0600 
Subject: Map Server Test Page up
Message-ID: <AEEF94298C10D311B5A000A024B359E1686CDF@imager9.spaceimaging.com>

The USGS (www.usgs.gov) provides a lot of vector data in DLG format. There
are several converters out there that will convert the SDTS format to shape.
The best method I have found is to use the SDTS2MIF and then the ESRI
MIF2SHAPE program. There's probably a better method, let me know if you find
it. A site called www.gisdatadepot.com claims to have over a terabyte of
vector and raster data from USGS and other sources. DRGs make a nice base
image. It's neatly organized and is freely downloadable.

ESRI also has a site called www.arcdataonline.com which has a mixture of low
cost and free data, mostly in shape format.

<blatant plug> Finally, Space Imaging sells 5 meter per pixed DOQs from IRS
(Indian Remote Sensing, not Internal Revenue Service :-) ) Last time I
checked they were going for about $250 each. Eventually we'll be selling
these online in MrSID format, but I'm not sure what the release date for
that project will be. Data from our new 1 meter per pixel IKONOS bird will
be available in the next 60-90 days, but I doubt this will fall into the
"cheap" category.</blatant plug>

Happy hunting.

Trevor

-----Original Message-----
From: Thomas E. Burk [mailto:teb at mallit.fr.umn.edu]
Sent: Wednesday, October 06, 1999 9:14 AM
To: mapserver-users at lists.gis.umn.edu
Subject: fwd: Map Server Test Page up


Forwarded with permission of the author.

Tom

------------- Begin Forwarded Message -------------


From pgallen at randomlogic.com  Mon Oct  4 22:54:32 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 04 Oct 1999 22:54:32 -0700
Subject: Map Server Test Page up
Message-ID: <mailman.151.1728305905.23453.mapserver-users@lists.osgeo.org>

I have fixed my firewall and my demo pages should now be available to all
(I've
had a couple people tell me they were successful at browsing the pages).
Once
again the URL is:

24.5.14.144
http://cx570538-a.elcjn1.sdca.home.com

If you're wondering why it's a strange URL or IP: I'm on a cable modem
(1MB/sec.
connection) using IP masquerading and forwarding on Red Hat Linux 6.0. Even
though I have 3 legitimate web sites, none of them are located on my home
machine and doing it this way gave me a free and easy method for testing
without
having to ftp changes all the time.

Question:

Is there a resource for up to date, low cost (or free) map databases? One of
the
issues my company has is the availability of current map data for the United
States. MapQuest provides it for $20,000/year (and an additional $50,000 for
the
map server!). If I can find a low cost solution, that'd be great.

PGA
------------- End Forwarded Message -------------


From pgallen at randomlogic.com  Wed Oct  6 11:17:58 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Wed, 06 Oct 1999 11:17:58 -0700
Subject: Map Server Test Page up
References: <199910061512.KAA27466@mallit.fr.umn.edu>
Message-ID: <37FB9256.5D94F291@randomlogic.com>

"Thomas E. Burk" wrote:

> [SNIP]
>
> You are subscribed. You've should have received some mail now. Thanks for
> allowing me to post your earlier message. You had been using <mapserver-info> to
> ask questions rather than <mapserver-users>.
>

Thanks. Yes, I have received a few messages since last night.

>
> Last night I added
> > queries to the San Diego and Philadelphia maps for lakes, streams and rivers,
> some
> > landmarks, and key geographic locations. I thinking about recompiling it on my
> > "legitimate" web server - a Sparc running FreeBSD - so I can keep it up for a
> long
> > term period as well as free up my machine for further development.
> >
> > BTW, I know how to code a router - for determining driving directions from
> point A
> > to point B - now it's just a matter of doing it. I've suggested to the local
> Linux
> > user group (Kernel Panic Linux User Group, KPLUG, http://www.kernel-panic.com)
> that
> > we make it a programing project in the programmers sub-group. No response yet.
> >
>
> Your work is very interesting to us. Thanks for your interest in MapServer. I
> hope the router project goes ahead. Keep us posted as you feel appropriate.
>

Thanks and no problem. I have done a few more improvements to the San Diego map:
Major Freeways, Minor Freeways, on/off ramps, Highways, Primary surface Streets, and
Secondary Surface are all different colors. I have also adjusted the scale at which
they become visible and added the query feature to them. It looks as though
MapServer will do everything we need aside from the routing, but routing should be a
separate package anyway, so that works out just fine. Now all we need to do is find
some good, inexpensive map data that also includes one-way street data (which
streets are one-way and which direction they go).

PGA



From ender at titan.lab.csuchico.edu  Thu Oct  7 10:10:41 1999
From: ender at titan.lab.csuchico.edu (ender at titan.lab.csuchico.edu)
Date: Thu, 7 Oct 1999 10:10:41 -0700 (PDT)
Subject: query
Message-ID: <Pine.LNX.3.96.991007100153.1461A-300000@titan.lab.csuchico.edu>


To all:

I have been attemping to perform a query using mapserver for some time
now, and I have not been successful.  I have scoured over all of the
documentation, as well as the samples, and have not been able to determine
what I am missing.

I always get the error:
An error has occurred in msQueryUsingPoint(): Search returned no results. 

Could somebody tell me exactly what this error means?  Has mapserver found
any layers to query?  How do you find out?

Attached is my .map and template file

-Aaron

-------------- next part --------------
<html>
<head>
<title>Compendium of Water Quality Investigations</title>
</head>
<body bgcolor=#ffffff>
<table width=100%>
<tr>
<td width=100% colspan=3 bgcolor="#585a87" cellspacing=0 cellpadding=0>
<b>Municipal Water Quality Investigations Program</b><br>
1997 Compendium of Water Quality Investigations<br>
Sampling Sites of the Sacramento River Watershed<br>
<br>
<font size=+1><b>Central Area</b></font><br>
<br>
</td></tr>
<form method=GET action="/cgi-bin/mapserv">
<tr>
<td align=center valign=top width=110 bgcolor="#585a87">
<a href="http://titan.lab.csuchico.edu/compendium/pages/northern.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_northern_area.jpg" border=0 alt="Northern Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/northeast.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_northeast_area.jpg" border=0 alt="Northeastern Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/central.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_central_area.jpg" border=0 alt="Central Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/eastern.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_eastern_area.jpg" border=0 alt="Eastern Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/southern.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_southern_area.jpg" border=0 alt="Southern Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/delta.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_delta_area.jpg" border=0 alt="Delta Area"></a>
<a href="http://titan.lab.csuchico.edu/compendium/pages/sfbay.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/b_sfbay_area.jpg" border=0 alt="SF Bay Area"></a>
</td>
<td valign=top>
<b>Programs conducting sampling within this area:</b><br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/F4.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/F4.jpg" border=0></a> - U.S. Geological Survey<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/F7.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/F7.jpg" border=0></a> - U.S. Geological Survey<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/L1.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/L1.jpg" border=0></a> - California State University, Chico<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/L19.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/L19.jpg" border=0></a> - Sand and Salt Creek Watershed Project<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/L20.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/L20.jpg" border=0></a> - Sacramento River Watershed Program<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S11.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S11.jpg" border=0></a> - Department of Water Resources<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S18.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S18.jpg" border=0></a> - <br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S19.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S19.jpg" border=0></a> - <br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S25.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S25.jpg" border=0></a> - State Water Resources Control Board<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S3.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S3.jpg" border=0></a> - Department of Fish and Game<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S7.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S7.jpg" border=0></a> - Department of Water Resources<br>
<a href="http://titan.lab.csuchico.edu/compendium/pages/S8.html"><img src="http://titan.lab.csuchico.edu/compendium/graphics/S8.jpg" border=0></a> - Department of Water Resources<br>
<td>
<input name=img type=image src="[img]" width= height= border=0><p>
Navigate:
<select name=zoomdir>
  <option value=1 [zoomdir_1_select]>Zoom In
  <option value=0 [zoomdir_0_select]>Pan
  <option value=-1 [zoomdir_-1_select]>Zoom Out
</select>
Operation:
<select name=mode>
  <option value=browse checked>Browse map
  <option value=query>Pick Sampling Site
</select>
</td></tr>
</table>

<input type=hidden name=layers value="[layers]">
<input type=hidden name=imgext value="[mapext]">
<input type=hidden name=map value="[map]">
<input type=hidden name=zoomsize value="2">
</form>
</body></html>
-------------- next part --------------
#
# Start of map file
#
NAME c
STATUS ON
SIZE 400 400
SHADESET /usr/local/share/ms/ms_symbols/symbols/shade.sym
MARKERSET /usr/local/share/ms/ms_symbols/symbols/marker.sym
LINESET /usr/local/share/ms/ms_symbols/symbols/line.sym
EXTENT -251552 65413.4 -119210 272783
UNITS METERS
SHAPEPATH "/usr/local/apache/htdocs/compendium/data"
IMAGECOLOR 255 255 255
transparent on
# LABELOVERLAP FALSE

#
# web interface
#
WEB
  template /usr/local/apache/htdocs/compendium/maps/central.html
END
#
#
# Start of reference map
#
REFERENCE
  IMAGE "/usr/local/apache/htdocs/compendium/graphics/reference_base.gif"
  EXTENT -373896 -604675 540167 449854
  STATUS ON
  COLOR -1 -1 -1
  OUTLINECOLOR 255 0 0
END

#
# Start of legend
#
LEGEND
  KEYSIZE 18 12
  LABEL
    TYPE BITMAP
    SIZE MEDIUM
    COLOR 0 0 89
  END
  STATUS ON
END

#
# Start of scalebar
#
SCALEBAR
  IMAGECOLOR 255 255 255
  LABEL
    COLOR 0 0 0
    SIZE SMALL
    ANTIALIAS
  END
  SIZE 350 5
  COLOR 255 255 255
  BACKGROUNDCOLOR 0 0 0
  OUTLINECOLOR 0 0 0
  UNITS KILOMETERS
  INTERVALS 5
  STATUS ON
END


#
# Start of layer definitions
#

# county lines

LAYER
  NAME county
  TYPE POLYGON
  STATUS DEFAULT
  DATA countys
  LABELITEM county_nam
  CLASS
    SYMBOL 0
    OUTLINECOLOR 230 230 230
    COLOR 255 255 255
    BACKGROUNDCOLOR 255 255 255
    LABEL
      SIZE SMALL
      TYPE truetype
      font Arial
      BUFFER 2
      COLOR 170 170 170
      POSITION cc
    END
  END
END # county layer

# hydro

LAYER
  NAME hydro
  TYPE LINE
  STATUS DEFAULT
  DATA central_hydro
  TOLERANCE 5
  LABELITEM PNAME
  CLASS
    COLOR  140 169 190 
    LABEL
     SIZE TINY
     COLOR  81 131 183 
    END
  END
END # hydro Layer

# lakes

LAYER
  NAME lakes
  TYPE POLYGON
  STATUS DEFAULT
  DATA central_lakes
  TOLERANCE 3
  CLASS
    COLOR  81 131 183 
  END
END # lakes Layer

# roads

LAYER
  NAME roads
  TYPE line
  status on
  data roads
  tolerance 5
  class
    color 168 155 143
  end
end # roads

LAYER
  NAME sites
  TYPE POINT
  STATUS ON
  DATA central_sites
  LABELITEM "NAME"
  CLASSITEM "ORG_TYPE"
  CLASS
    SYMBOL 0
    COLOR 255 0 0
    EXPRESSION "F"
    LABEL
      POSITION ul
      COLOR 255 0 0
      backgroundcolor  255 255 255 
      buffer 2
    END
  END
  CLASS
    SYMBOL 0
    COLOR 13 130 0
    EXPRESSION "S"
    LABEL
      POSITION ul
      COLOR 13 130 0
      backgroundcolor  255 255 255 
      buffer 2
    END
  END
  CLASS
    SYMBOL 0
    COLOR 211 29 111
    EXPRESSION "L"
    LABEL
      POSITION ul
      COLOR 128 0 128
      backgroundcolor  255 255 255 
      buffer 2
    END
  END

  QUERYITEM "NAME"
  QUERY
    EXPRESSION "^-"
    TEMPLATE "site.html"
  END
END # sites layer

# cities

LAYER
  NAME cities
  TYPE POINT
  STATUS on
  DATA central_cities
  TOLERANCE 10

  LABELITEM "NAME"

  CLASS
    SYMBOL 0
    SIZE 4
    LABEL
      COLOR  0 0 0 
      POSITION ul
      PARTIALS off
      backgroundcolor  200 200 200 
      buffer 5
    END
  END
END # cities layer
END # end of mapfile

From pgallen at randomlogic.com  Thu Oct  7 12:58:49 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Thu, 07 Oct 1999 12:58:49 -0700
Subject: query
References: <Pine.LNX.3.96.991007100153.1461A-300000@titan.lab.csuchico.edu>
Message-ID: <37FCFB79.3A958F79@randomlogic.com>

ender at titan.lab.csuchico.edu wrote:

> To all:
>
> I have been attemping to perform a query using mapserver for some time
> now, and I have not been successful.  I have scoured over all of the
> documentation, as well as the samples, and have not been able to determine
> what I am missing.
>
> I always get the error:
> An error has occurred in msQueryUsingPoint(): Search returned no results.
>
> Could somebody tell me exactly what this error means?  Has mapserver found
> any layers to query?  How do you find out?
>
> Attached is my .map and template file
>
> [BIG SNIP]

I had the same problem you do, and I believe the reason is the same as well. I'd forgotten to put the "TOLERANCE" line in my map file (see below). The TOLERANCE keyword sets the radius in pixels
from the point of the mouse click that a query will take place. The default is 0, meaning the mouse has to be on the exact point of the object to query it. In the case of a large area - a lake,
city region, etc. -  this is OK, but in the case of a point, it usually isn't. Setting a TOLERANCE of 5 - 10 (possibly more) should do the trick. Hope this helps.

PGA

>
>
> LAYER
>   NAME sites
>   TYPE POINT
>   STATUS ON
>   DATA central_sites

#   PGA: Added TOLERANCE

    TOLERANCE 10

#===================

>
>   LABELITEM "NAME"
>   CLASSITEM "ORG_TYPE"
>   CLASS
>     SYMBOL 0
>     COLOR 255 0 0
>     EXPRESSION "F"
>     LABEL
>       POSITION ul
>       COLOR 255 0 0
>       backgroundcolor  255 255 255
>       buffer 2
>     END
>   END
>   CLASS
>     SYMBOL 0
>     COLOR 13 130 0
>     EXPRESSION "S"
>     LABEL
>       POSITION ul
>       COLOR 13 130 0
>       backgroundcolor  255 255 255
>       buffer 2
>     END
>   END
>   CLASS
>     SYMBOL 0
>     COLOR 211 29 111
>     EXPRESSION "L"
>     LABEL
>       POSITION ul
>       COLOR 128 0 128
>       backgroundcolor  255 255 255
>       buffer 2
>     END
>   END
>
>   QUERYITEM "NAME"
>   QUERY
>     EXPRESSION "^-"
>     TEMPLATE "site.html"
>   END
> END # sites layer
>
> # cities
>
> LAYER
>   NAME cities
>   TYPE POINT
>   STATUS on
>   DATA central_cities
>   TOLERANCE 10
>
>   LABELITEM "NAME"
>
>   CLASS
>     SYMBOL 0
>     SIZE 4
>     LABEL
>       COLOR  0 0 0
>       POSITION ul
>       PARTIALS off
>       backgroundcolor  200 200 200
>       buffer 5
>     END
>   END
> END # cities layer
> END # end of mapfile






From steve.lime at dnr.state.mn.us  Thu Oct  7 14:58:40 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Thu, 07 Oct 1999 16:58:40 -0500
Subject: mapserver and database connectivity.
Message-ID: <s7fcd150.094@smtp.dnr.state.mn.us>

I'm wrapping mapserver with SWIG now. Going ok, SWIG ain't perfect but most things I want to do should work. The perl interface allows you to define arbitrary features (points, lines and polygons) and add them to a map composition. This way just about any data source that's accessible from perl can be added to the mapserv.

Should be doable from PHP as well although I'm not sure if there is a convientent tool like SWIG for extending it. I actually had a PHP module working a few years ago so I know it can be done. 

I don't think extending the C codebase is the way to go unless there's a decent library out there. Too much reinventing the wheel otherwise.

Steve

<<< Nathan Carr <ncarr at guideguide.com> 10/ 6  4:56a >>>
	I too am interested in this functionality. Have been looking at
wrapping
mapserver with SWIG, so I could use perl DBI to access database tables.
This seems a bit
confusing as there is quite a bit of libraries and structures to
include.
 
Alternatively would it be possible to extend the JOIN functionality for
joining DBF
files to join MySQL/PostGres type tables? 

Cheerio, nathan

> 
> Hi all,
> 
> Second this question.  We are looking at building similar functionality,
> but with PHP.  When we discover any info. re. this, we will post it.
> Perhaps Steve has some insight?
> 
> Cheers,
> Jim
> 
> Robert Chavez wrote:
> >
> > Greetings,
> >
> > I am wondering if it is possible to send information to mapserver from a
> > database (PostGresql).
> >
> >  I'm using the Pg perl module to build a web interface to my PostGres
> > database so users can query and display the data set.  The next logical step
> > would be to send the results to Mapserver somehow and plot them on my
> > basemap -- or perhaps create a tranparent gif of the coordinates that could
> > be overlayed on the base map.  Has anyone had any experience  with this sort
> > of functionality?
> >
> > Thanks,
> > Rob Chavez
> 
> --
> 
-- 
Nathan Carr  ncarr at guideguide.com



From pgallen at randomlogic.com  Thu Oct  7 17:57:18 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Thu, 07 Oct 1999 17:57:18 -0700
Subject: mapserver and database connectivity.
References: <s7fcd150.094@smtp.dnr.state.mn.us>
Message-ID: <37FD416D.2B2CC44D@randomlogic.com>

Stephen Lime wrote:

> I'm wrapping mapserver with SWIG now. Going ok, SWIG ain't perfect but most things I want to do should work. The perl interface allows you to define arbitrary features (points, lines and polygons) and add them to a map composition. This way just about any data source that's accessible from perl can be added to the mapserv.
>
> Should be doable from PHP as well although I'm not sure if there is a convientent tool like SWIG for extending it. I actually had a PHP module working a few years ago so I know it can be done.
>
> I don't think extending the C codebase is the way to go unless there's a decent library out there. Too much reinventing the wheel otherwise.
>

I'm very interested in this as it seems it would apply directly to what we want to do - insert GPS data points into a map. I could do this now by simply converting the data into data sets that mapserver can read and simply reading that data, but if there's another, easier method...

I know nothing about SWIG, not even what it stands for, but I am learning perl. With that, what is SWIG? (Scientific Wild Insane Guess?! :)

PGA






From cshorte2 at csc.com  Thu Oct  7 18:56:53 1999
From: cshorte2 at csc.com (cshorte2 at csc.com)
Date: Fri, 8 Oct 1999 11:56:53 +1000
Subject: mapserver and database connectivity.
Message-ID: <85256804.000B4E52.00@csc.com>



I assume they are refering to http://www.swig.org/

>From the homepage:

SWIG is a software development tool that connects programs written in C, C++,
and Objective-C
with a variety of high-level programming languages. SWIG is primarily used with
common
scripting languages such as Perl, Python, and Tcl/Tk, but has also been extended
 to include
languages such as Java, Eiffel, and Guile. SWIG is most commonly used to create
high-level
interpreted programming environments, systems integration, and as a tool for
building user
interfaces. SWIG may be freely used, distributed, and modified for commercial or
 noncommercial
use.

> I know nothing about SWIG, not even what it stands for,
> but I am learning perl. With that, what is SWIG? (Scientific Wild Insane
Guess?! :)




From pgallen at randomlogic.com  Fri Oct  8 13:15:28 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 08 Oct 1999 13:15:28 -0700
Subject: US Street Map
Message-ID: <37FE50E0.5C055B26@randomlogic.com>

I plan to put up a map of the entire US including every street, freeway, city,
county, etc. I have all the data, but I need to convert it into the proper
format. The data files are currently in TIGER/Line format as text files (see
http://www.perens.com or fpt://ftp.perens.com/pub/US-map for more info). Each
file is a separate TIGER/Line record type. I've heard there are a lot of
converter programs available. Is there a converter for this format that will put
the data in the format necessary for MapServer - to convert it into ESRI format
.shp, .shx, and .dbf files?

Thanks,

PGA

P.S. - If anyone is wanting to D/L the data, fair warning: it took me most of
the night on my 1.5MB/sec cable modem. It's 2.6GB of bzipped text files in 57
different directories. I'm glad I have NcFTP! :)



From steve.lime at dnr.state.mn.us  Fri Oct  8 13:17:06 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 08 Oct 1999 15:17:06 -0500
Subject: request for comments...
Message-ID: <s7fe0ba6.041@smtp.dnr.state.mn.us>

I'm in the midst of implementing a new feature for 3.3 called "query maps".
In a nutshell they allow you to use the results of a query to generate a map.
The constructor is very simple, you specify 3 parameters status, style and 
a hilite color (if necessary). I envision 3 styles of query maps:

NORMAL - with this style the results of the query would be used to define
a map extent and then the map would be draw as normal with no special
attention payed to the query polygons. This would allow for item/value
queries to generate a map in one step rather than the current two. Embedding
maps if you know nothing about coordinates would be much easier this way.

HILITE - the map would be drawn as normal, but features that are part of
the query result set would be drawn using the hilite color. Very ArcView-ish.

SELECTED - for queried layers, only those features that are part of the 
result set would be drawn. They would be rendered normally with no hilites.
Especially useful with item/value queries. Non-query layers would be drawn
as normal.

Note that queries are limited to shapefiles only. Tiled data is not queryable.

(Using perl and data from other sources will in effect give you query access
to those sources since drawing is done feature by feature.)

Anyway, I'd like to get comments from folks on this extension. Other ideas and
so on. I can envision some very sophisticated apps just using what I've outlined
above.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us



From cshorter at optusnet.com.au  Fri Oct  8 14:05:47 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Sat, 09 Oct 1999 07:05:47 +1000
Subject: request for comments...
References: <s7fe0ba6.041@smtp.dnr.state.mn.us>
Message-ID: <37FE5CAB.21541C0A@optusnet.com.au>

Yes, I can see myself using such a feature.  What I am working toward is an
interface to modify map shapes.  The way I see it working is:

Split my web page into windows.
Window 1, the map window, contains a map.  From this window, an operator can
select a feature (like a road), and all the details of that object are copied
to window 2.  The operator can also enter a new feature, or update an old
feature by entering multiple points on the map window.

Window 2, the feature details window, contains all the vertex and data about
the selected feature.  When the opertor has finished updating the feature,
they press the SUBMIT button and the details are sent to the database to be
updated.

Stephen Lime wrote:
> 
> I'm in the midst of implementing a new feature for 3.3 called "query maps".
> In a nutshell they allow you to use the results of a query to generate a map.
> The constructor is very simple, you specify 3 parameters status, style and
> a hilite color (if necessary). I envision 3 styles of query maps:
> 
> NORMAL - with this style the results of the query would be used to define
> a map extent and then the map would be draw as normal with no special
> attention payed to the query polygons. This would allow for item/value
> queries to generate a map in one step rather than the current two. Embedding
> maps if you know nothing about coordinates would be much easier this way.
> 
> HILITE - the map would be drawn as normal, but features that are part of
> the query result set would be drawn using the hilite color. Very ArcView-ish.
> 
> SELECTED - for queried layers, only those features that are part of the
> result set would be drawn. They would be rendered normally with no hilites.
> Especially useful with item/value queries. Non-query layers would be drawn
> as normal.
> 
> Note that queries are limited to shapefiles only. Tiled data is not queryable.
> 
> (Using perl and data from other sources will in effect give you query access
> to those sources since drawing is done feature by feature.)
> 
> Anyway, I'd like to get comments from folks on this extension. Other ideas and
> so on. I can envision some very sophisticated apps just using what I've outlined
> above.
> 
> Steve
> 
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
> 
> (651) 297-2937
> steve.lime at dnr.state.mn.us


From pcameron at gis.umn.edu  Fri Oct  8 14:46:32 1999
From: pcameron at gis.umn.edu (Paul Cameron)
Date: Fri, 8 Oct 1999 16:46:32 -0500 (CDT)
Subject: request for comments...
In-Reply-To: <s7fe0ba6.041@smtp.dnr.state.mn.us>
Message-ID: <Pine.SGI.3.91.991008162548.943A-100000@center>

I am very much in favor of seeing something like this implemented. I'm 
got an app running now that sounds very similar to the HILITE mode but 
its a very convoluted process that I'd love to be able to clean up. I've 
got a couple comments/questions on them below.

On Fri, 8 Oct 1999, Stephen Lime wrote:

> NORMAL - with this style the results of the query would be used to define
> a map extent and then the map would be draw as normal with no special
> attention payed to the query polygons. This would allow for item/value
> queries to generate a map in one step rather than the current two. Embedding
> maps if you know nothing about coordinates would be much easier this way.

If I understand this correctly than this will remove the problem of 
having to know the mapextent of a polygon that we want to zoom to. Will 
we be able to specify a different mapfile for the new map that will be 
created? This would be very useful for nesting interfaces together.

> 
> HILITE - the map would be drawn as normal, but features that are part of
> the query result set would be drawn using the hilite color. Very ArcView-ish.

I envision this working by passing an array of selected features to 
mapserv which would color them in for me. Will we be able to specify the 
color used in the hilite and can we make the highlight color transparent? 

This one might be stretching but for the application I've got going right 
now I need to break the highlighted features into different classes. 
Could multiple classes be supported, possibily using hashes in perl?

I'm also curious how this will be integrated into mapscript. I don't 
know how mapscript will be work but I could see this being closely tied 
to using mapscript. 

paul

> 
> SELECTED - for queried layers, only those features that are part of the 
> result set would be drawn. They would be rendered normally with no hilites.
> Especially useful with item/value queries. Non-query layers would be drawn
> as normal.
> 
> Note that queries are limited to shapefiles only. Tiled data is not queryable.
> 
> (Using perl and data from other sources will in effect give you query access
> to those sources since drawing is done feature by feature.)
> 
> Anyway, I'd like to get comments from folks on this extension. Other ideas and
> so on. I can envision some very sophisticated apps just using what I've outlined
> above.
> 
> Steve
> 
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
> 
> (651) 297-2937
> steve.lime at dnr.state.mn.us
> 




From sphillip at mail.esrp.org  Fri Oct  8 15:04:40 1999
From: sphillip at mail.esrp.org (Scott Phillips)
Date: Fri, 08 Oct 1999 15:04:40 PDT
Subject: US Street Map (fwd)
Message-ID: <199910082204.PAA26337@mail.esrp.org>

Paul,

There is an converter for windows that costs about $100.00 called TGR2SHP.
I used it to convert the State of California set to shape files and it 
worked great.  It lets you do the conversions straight from the .ZIP files
and you can do them in batch.   See http://www.gistools.com if you want
more info.

Scott

Scott Phillips
scott at phillipsgis.com

> 
> I plan to put up a map of the entire US including every street, freeway, city,
> county, etc. I have all the data, but I need to convert it into the proper
> format. The data files are currently in TIGER/Line format as text files (see
> http://www.perens.com or fpt://ftp.perens.com/pub/US-map for more info). Each
> file is a separate TIGER/Line record type. I've heard there are a lot of
> converter programs available. Is there a converter for this format that will put
> the data in the format necessary for MapServer - to convert it into ESRI format
> .shp, .shx, and .dbf files?
> 
> Thanks,
> 
> PGA
> 
> P.S. - If anyone is wanting to D/L the data, fair warning: it took me most of
> the night on my 1.5MB/sec cable modem. It's 2.6GB of bzipped text files in 57
> different directories. I'm glad I have NcFTP! :)
> 
> 



From pgallen at randomlogic.com  Fri Oct  8 15:21:39 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 08 Oct 1999 15:21:39 -0700
Subject: US Street Map (fwd)
References: <199910082204.PAA26337@mail.esrp.org>
Message-ID: <37FE6E73.D395CFBD@randomlogic.com>

Scott Phillips wrote:

> Paul,
>
> There is an converter for windows that costs about $100.00 called TGR2SHP.
> I used it to convert the State of California set to shape files and it
> worked great.  It lets you do the conversions straight from the .ZIP files
> and you can do them in batch.   See http://www.gistools.com if you want
> more info.
>

This will do the trick but is there a UNIX/Linux equivalent. (I prefer UNIX/Linux over
the alternative, but I can use TGR2SHP if I must).

Thanks,

PGA





From nhv at cape.com  Fri Oct  8 16:57:37 1999
From: nhv at cape.com (Norman Vine)
Date: Fri, 8 Oct 1999 19:57:37 -0400
Subject: US Street Map
In-Reply-To: <37FE50E0.5C055B26@randomlogic.com>
Message-ID: <000001bf11e9$024b3a80$1637ba8c@nhv>

Paul G. Allen writes:

>Is there a converter for this 
>format that will put the data in the format necessary for MapServer
>to convert it into ESRI format .shp, .shx, and .dbf files?
>

Alan Crosswell has GPL'd 'C' source 
for making maps from Tiger 94-95 data at
http://www.users.cloud9.net/~alan/ham/aprs/tig2aprs.zip

that could be used as input to

Frank Warmerdam's Shapelib package
http://members.home.com/warmerda

To make the shapefiles.

Hopefully the format did not changed that much from 95 to 97 

Good Luck

Norman Vine


From pnaciona at gis.umn.edu  Fri Oct  8 11:02:37 1999
From: pnaciona at gis.umn.edu (Pericles S. Nacionales)
Date: Fri, 8 Oct 1999 19:02:37 +0100
Subject: request for comments...
References: <s7fe0ba6.041@smtp.dnr.state.mn.us>
Message-ID: <002c01bf11b7$4e930ad0$9a496580@gis.umn.edu>

Steve,

  This is perfect for one of our TerraSIP apps--the TM Availability tool.
We can generate the map with the query results (the dates LandSat 5 or 7
passes the Upper Great Lakes), have the scene they selected highlighted, and
have a GOES image alongside for users to see if it's a clear or cloudy day.
All in one query...  this is good.  I assume you can then requery from the
new map?

Thanks,
Perry

PS:  Terrasip prototype MapServer apps are at
http://terrasip.gis.umn.edu/projects


Pericles S. Nacionales
Research Assistant
Forest Resources Dept.
University of Minnesota

612 625 5765
pnaciona at gis.umn.edu



----- Original Message -----
From: Stephen Lime <steve.lime at dnr.state.mn.us>
To: <mapserver-users at lists.gis.umn.edu>
Sent: Friday, October 08, 1999 9:17 PM
Subject: request for comments...


> I'm in the midst of implementing a new feature for 3.3 called "query
maps".
> In a nutshell they allow you to use the results of a query to generate a
map.
> The constructor is very simple, you specify 3 parameters status, style and
> a hilite color (if necessary). I envision 3 styles of query maps:
>
> NORMAL - with this style the results of the query would be used to define
> a map extent and then the map would be draw as normal with no special
> attention payed to the query polygons. This would allow for item/value
> queries to generate a map in one step rather than the current two.
Embedding
> maps if you know nothing about coordinates would be much easier this way.
>
> HILITE - the map would be drawn as normal, but features that are part of
> the query result set would be drawn using the hilite color. Very
ArcView-ish.
>
> SELECTED - for queried layers, only those features that are part of the
> result set would be drawn. They would be rendered normally with no
hilites.
> Especially useful with item/value queries. Non-query layers would be drawn
> as normal.
>
> Note that queries are limited to shapefiles only. Tiled data is not
queryable.
>
> (Using perl and data from other sources will in effect give you query
access
> to those sources since drawing is done feature by feature.)
>
> Anyway, I'd like to get comments from folks on this extension. Other ideas
and
> so on. I can envision some very sophisticated apps just using what I've
outlined
> above.
>
> Steve
>
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
>
> (651) 297-2937
> steve.lime at dnr.state.mn.us
>



From steve.lime at dnr.state.mn.us  Fri Oct  8 20:03:59 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 08 Oct 1999 22:03:59 -0500
Subject: request for comments...
Message-ID: <s7fe6a67.071@smtp.dnr.state.mn.us>

Think of query maps as regular maps. Instead of pan/zoom you do a query of some sort. They don't make sense in all cases but should pretty helpful in others. The output template could easily be another mapserver interface. Preserving queries may be tough though.

Steve 

<<< "Pericles S. Nacionales" <pnaciona at lists.gis.umn.edu> 10/ 8  7:02p >>>
Steve,

  This is perfect for one of our TerraSIP apps--the TM Availability tool.
We can generate the map with the query results (the dates LandSat 5 or 7
passes the Upper Great Lakes), have the scene they selected highlighted, and
have a GOES image alongside for users to see if it's a clear or cloudy day.
All in one query...  this is good.  I assume you can then requery from the
new map?

Thanks,
Perry

PS:  Terrasip prototype MapServer apps are at
http://terrasip.gis.umn.edu/projects


Pericles S. Nacionales
Research Assistant
Forest Resources Dept.
University of Minnesota

612 625 5765
pnaciona at gis.umn.edu



----- Original Message -----
From: Stephen Lime <steve.lime at dnr.state.mn.us>
To: <mapserver-users at lists.gis.umn.edu>
Sent: Friday, October 08, 1999 9:17 PM
Subject: request for comments...


> I'm in the midst of implementing a new feature for 3.3 called "query
maps".
> In a nutshell they allow you to use the results of a query to generate a
map.
> The constructor is very simple, you specify 3 parameters status, style and
> a hilite color (if necessary). I envision 3 styles of query maps:
>
> NORMAL - with this style the results of the query would be used to define
> a map extent and then the map would be draw as normal with no special
> attention payed to the query polygons. This would allow for item/value
> queries to generate a map in one step rather than the current two.
Embedding
> maps if you know nothing about coordinates would be much easier this way.
>
> HILITE - the map would be drawn as normal, but features that are part of
> the query result set would be drawn using the hilite color. Very
ArcView-ish.
>
> SELECTED - for queried layers, only those features that are part of the
> result set would be drawn. They would be rendered normally with no
hilites.
> Especially useful with item/value queries. Non-query layers would be drawn
> as normal.
>
> Note that queries are limited to shapefiles only. Tiled data is not
queryable.
>
> (Using perl and data from other sources will in effect give you query
access
> to those sources since drawing is done feature by feature.)
>
> Anyway, I'd like to get comments from folks on this extension. Other ideas
and
> so on. I can envision some very sophisticated apps just using what I've
outlined
> above.
>
> Steve
>
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
>
> (651) 297-2937
> steve.lime at dnr.state.mn.us
>




From TBanister at SpaceImaging.com  Sat Oct  9 10:09:14 1999
From: TBanister at SpaceImaging.com (Banister Trevor)
Date: Sat, 9 Oct 1999 11:09:14 -0600 
Subject: US Street Map
Message-ID: <AEEF94298C10D311B5A000A024B359E1686CF5@imager9.spaceimaging.com>

If you have ArcView, there is a free Avenue script at:
http://gis.esri.com/arcscripts/details.cfm?CFGRIDKEY=1270853054

-----Original Message-----
From: Paul G. Allen [mailto:pgallen at randomlogic.com]
Sent: Friday, October 08, 1999 2:15 PM
To: MapServer Users List
Subject: US Street Map


I plan to put up a map of the entire US including every street, freeway,
city,
county, etc. I have all the data, but I need to convert it into the proper
format. The data files are currently in TIGER/Line format as text files (see
http://www.perens.com or fpt://ftp.perens.com/pub/US-map for more info).
Each
file is a separate TIGER/Line record type. I've heard there are a lot of
converter programs available. Is there a converter for this format that will
put
the data in the format necessary for MapServer - to convert it into ESRI
format
.shp, .shx, and .dbf files?

Thanks,

PGA

P.S. - If anyone is wanting to D/L the data, fair warning: it took me most
of
the night on my 1.5MB/sec cable modem. It's 2.6GB of bzipped text files in
57
different directories. I'm glad I have NcFTP! :)


From imap at chesapeake.net  Sat Oct  9 16:45:14 1999
From: imap at chesapeake.net (imap at chesapeake.net)
Date: Sat, 09 Oct 1999 19:45:14 -0400
Subject: US Street Map
References: <AEEF94298C10D311B5A000A024B359E1686CF5@imager9.spaceimaging.com>
Message-ID: <37FFD38A.1C49BF43@chesapeake.net>


Banister (very cool name),

That TGR2SHP is the way to go (www.gistools.com) on this one.
It does ENTIRE cdroms in batches...  (I had requested this
feature about a year ago, and he did it).  There is so much
data, you have to automate as much as possible.

I work with this Tiger data all the time and that product
is just the ticket...  The only trick is to post-process
the files to get them in the correct projection.  I wrote
my own projection tool for doing this... using PROJ4
which can be found on the net as well... 

ftp://kai.er.usgs.gov/pub/PROJ.4/

Regards,

Chris Stuber (mapsurfer)


Banister Trevor wrote:
> 
> If you have ArcView, there is a free Avenue script at:
> http://gis.esri.com/arcscripts/details.cfm?CFGRIDKEY=1270853054
> 
> -----Original Message-----
> From: Paul G. Allen [mailto:pgallen at randomlogic.com]
> Sent: Friday, October 08, 1999 2:15 PM
> To: MapServer Users List
> Subject: US Street Map
> 
> I plan to put up a map of the entire US including every street, freeway,
> city,
> county, etc. I have all the data, but I need to convert it into the proper
> format. The data files are currently in TIGER/Line format as text files (see
> http://www.perens.com or fpt://ftp.perens.com/pub/US-map for more info).
> Each
> file is a separate TIGER/Line record type. I've heard there are a lot of
> converter programs available. Is there a converter for this format that will
> put
> the data in the format necessary for MapServer - to convert it into ESRI
> format
> .shp, .shx, and .dbf files?
> 
> Thanks,
> 
> PGA
> 
> P.S. - If anyone is wanting to D/L the data, fair warning: it took me most
> of
> the night on my 1.5MB/sec cable modem. It's 2.6GB of bzipped text files in
> 57
> different directories. I'm glad I have NcFTP! :)

-- 

Chris Stuber (mapsurfer)
287 Southland Ct
Dunkirk, MD  20754


From cshorter at optusnet.com.au  Sun Oct 10 03:14:45 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Sun, 10 Oct 1999 20:14:45 +1000
Subject: Updating Shapefiles, and data integrity
Message-ID: <38006715.1C3C48C0@optusnet.com.au>



-------- Original Message --------
Subject: Re: request for comments...
Date: Sun, 10 Oct 1999 19:16:24 +1000
From: Cameron Shorter <cshorter at optusnet.com.au>
Reply-To: camerons at cat.org.au
To: Stephen Lime <steve.lime at dnr.state.mn.us>
References: <s7ff017e.048 at smtp.dnr.state.mn.us>



Stephen Lime wrote:
> 
> Maintaining data integrity is going to be a big issue. I was at our state GIS conference and got to chat with Jack Dangermond from ESRI about the MapServer and their new ArcIMS product. Seems they're having trouble with this editing stuff. Shapefiles just aren't a transactional environment so unless you can assure yourself of single user access there's always the potential for multiple concurrent edits. Then there's the issue of quality control. I think the solution needs to offer immediate update and delayed update. ArcIMS, as I understand it, caches updates until an operator on the server site commits the edits to the main database. This operator could be a cron process I suppose that could handle locking while
> edits are processed. I think this may be a good approach as you could do some simple transaction management- review, edit and delete, once the initial work was done. Edits could be stored in a shapefile along with attributes and enough additional information to commit the shape - source shapefile, shape (or is a new one), type of edit (replace, attribute change) etc.
> 
> Anyway, just my thoughts...
> 
> Steve

I'm glad to hear I'm not the only one having problems with updating
shapefiles. :)

>From looking at the shapefile definition paper, you can see that there is an
index file .SHX which points to a .SHP file which has variable length records.

There are a few problems that I can see.  Please verify if any of these are
correct or not.
1. Deleting an old object.  I think this can be handled by setting the
shapetype to a NULL shape.

2. Increasing the number of vertices of a shape, and hence increasing the
record size.  I think the best way to handle this is to remove the old shape
by setting its shapetype to NULL, and to add a new shape to the end of the
.SHP file.  The pointer in the .SHX file will now have to be redirected to the
end of the .SHP file.  This now means that the order of the .SHP file and the
.SHX file will not match, which will reduce query speeds, so periodically the
datafiles would need to be rebuilt.

3. There is an issue with the .SHP file and .SHX file becoming out of sync. 
Basically, when a shape is updated, first the .SHP file will need to be
updated, and some time later the .SHX file will be updated.  There is a window
of opportunity where the files will be out of sync.  I was planning to address
this by either putting in a lock file, or changing read/write permissions to
lock the files while the database is out of sync.
This means that some reads of the database will fail because the database is
updating.

4. I'm not sure what the best way is to link into a SQL database.  If the
shapefile is only added to, then the best way to reference an object is by
using the index in the .SHX file.  However, if you delete an object, should
you rebuild the .SHX file?  This will keep the index file from blowing out,
but all the indexes will change and hence the  SQL database will reference the
wrong indices.

Happy for any advice.

Cameron.


From vrraju at gis.umn.edu  Mon Oct 11 08:38:09 1999
From: vrraju at gis.umn.edu (Vatsavai Ranga Raju)
Date: Mon, 11 Oct 1999 10:38:09 -0500
Subject: MapServer List Archive
References: <s7ed4b51.060@smtp.dnr.state.mn.us> <37EDA085.8BFD550B@optusnet.com.au> <37EDAABC.EDDC798@gis.umn.edu>
Message-ID: <38020461.1799DBCE@gis.umn.edu>

To mapserver-users list members

To improve the usefulness of archives, we have decided to update the html archives
every mid night. You can access the html archives from the following URL:
http://mapserver.gis.umn.edu/mapserver-users-list-archives/index.html

Raju
----

Vatsavai Ranga Raju wrote:

> Hi Cameron
>
> Sure, email interface (i.e - get list-name file-name from majordomo) is not a convenient
> method to go through mail archives. But there are utilities (like mhonarc) which will convert
> email list to html files. Initially there is very little traffic for mapserver-users lists, so we
> thought monthly archives are sufficient (for html interface, but users can get up-to-date
> archive files through email interface). Ofcourse we can archive them at frequent intervals,
> if there is sufficient traffic and need.
>
> Raju.
> ----
>
> Cameron Shorter wrote:
>
> > I'd be really impressed if the archive could be made more recent.  Month old
> > archives often arn't good enough for me, and I find the email interface messy.
> >
> > Cameron
> >
> > Stephen Lime wrote:
> > >
> > > Yup, the University maintains a web version. It's a bit out of date *hint*, but I know they plan on keeping it up.



From pgallen at randomlogic.com  Mon Oct 11 15:24:38 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 11 Oct 1999 15:24:38 -0700
Subject: Tiling maps
Message-ID: <380263A6.B00A5039@randomlogic.com>

I am ready to tile maps but I'm not sure how to do this. I have fairly good data
for the entire US, separated by county. What I want to do is to be able to
display a map of any size from a single county to several counties, possibly an
entire state. My question is, how would I display two adjacent counties as a
single map?

PGA



From steve.lime at dnr.state.mn.us  Tue Oct 12 07:25:07 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 12 Oct 1999 09:25:07 -0500
Subject: Tiling maps
Message-ID: <s802fe98.014@smtp.dnr.state.mn.us>

That's where the tiling comes in. You need a county index shapefile. The
index is just a normal shapefile with a attribute called "location" that contains
the path to the directory that contains data for that tile. The path can be a
full path (eg. starts with /) or it can be relative to the location of the mapfile.
(I think that's correct) To enable tiling you simply use the INDEX keyword
(followed by the index shapefile name) in a layer section.

The MapServer uses the index to figure out which tiles to process. Then, each
one is processed in turn as if it was a single shapefile. Only drag is that tiled
layers are not searchable yet.

A number of MapServer users are using this feature and are in a position to
comment about specific problems you might run in to.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

>>> "Paul G. Allen" <pgallen at randomlogic.com> 10/11 5:24 PM >>>
I am ready to tile maps but I'm not sure how to do this. I have fairly good data
for the entire US, separated by county. What I want to do is to be able to
display a map of any size from a single county to several counties, possibly an
entire state. My question is, how would I display two adjacent counties as a
single map?

PGA




From bfraser at geoanalytic.ab.ca  Tue Oct 12 07:40:49 1999
From: bfraser at geoanalytic.ab.ca (bfraser)
Date: Tue, 12 Oct 1999 08:40:49 -0600
Subject: Tiling maps
References: <380263A6.B00A5039@randomlogic.com>
Message-ID: <001801bf14bf$c6b0fc60$390002c0@servo>

Paul,

  As long as you don't want to query the data in the tiles, you
can use the tiling feature in MapServer.

1.  Create a shapefile containing the outlines of the counties
2.  Create a column in the shapefile with the name of LOCATION
3.  For each county outline shape in the shapefile, enter the filename
     of the county shapefile in the LOCATION column (leave off the
     .shp )
4.  Edit the "map" file.   Add an entry for TILE at the top.   This
     contains the path to the directory containing county files.  It is
     prepended to the name in the LOCATION column to build the
     full path to the shapefiles.

    Add a  LAYER entry with an INDEX entry.  The index is the name
    of the shapefile you created in step 1.

A couple of notes about tiles:
    1. The tile index shapefile and the "data" shapefiles must be in
        the same coordinate system.
    2. Mapserver will not query data in tiled files.

Brent


----- Original Message -----
From: Paul G. Allen <pgallen at randomlogic.com>
To: MapServer Users List <mapserver-users at lists.gis.umn.edu>
Sent: Monday, October 11, 1999 4:24 PM
Subject: Tiling maps


> I am ready to tile maps but I'm not sure how to do this. I have fairly
good data
> for the entire US, separated by county. What I want to do is to be able to
> display a map of any size from a single county to several counties,
possibly an
> entire state. My question is, how would I display two adjacent counties as
a
> single map?
>
> PGA
>
>



From bfraser at geoanalytic.ab.ca  Tue Oct 12 09:00:08 1999
From: bfraser at geoanalytic.ab.ca (bfraser)
Date: Tue, 12 Oct 1999 10:00:08 -0600
Subject: Updating Shapefiles, and data integrity
References: <38006715.1C3C48C0@optusnet.com.au>
Message-ID: <006c01bf14ca$e68f3f00$390002c0@servo>

see my comments below...

Brent Fraser


----- Original Message -----
From: Cameron Shorter <cshorter at optusnet.com.au>
To: mapserver <mapserver-users at lists.gis.umn.edu>
Sent: Sunday, October 10, 1999 4:14 AM
Subject: Updating Shapefiles, and data integrity


>
>
> -------- Original Message --------
> Subject: Re: request for comments...
> Date: Sun, 10 Oct 1999 19:16:24 +1000
> From: Cameron Shorter <cshorter at optusnet.com.au>
> Reply-To: camerons at cat.org.au
> To: Stephen Lime <steve.lime at dnr.state.mn.us>
> References: <s7ff017e.048 at smtp.dnr.state.mn.us>
>
>
>
> Stephen Lime wrote:
> >
> > Maintaining data integrity is going to be a big issue. I was at our
state GIS conference and got to chat with Jack Dangermond from ESRI about
the MapServer and their new ArcIMS product. Seems they're having trouble
with this editing stuff. Shapefiles just aren't a transactional environment
so unless you can assure yourself of single user access there's always the
potential for multiple concurrent edits. Then there's the issue of quality
control. I think the solution needs to offer immediate update and delayed
update. ArcIMS, as I understand it, caches updates until an operator on the
server site commits the edits to the main database. This operator could be a
cron process I suppose that could handle locking while
> > edits are processed. I think this may be a good approach as you could do
some simple transaction management- review, edit and delete, once the
initial work was done. Edits could be stored in a shapefile along with
attributes and enough additional information to commit the shape - source
shapefile, shape (or is a new one), type of edit (replace, attribute change)
etc.
> >
> > Anyway, just my thoughts...
> >
> > Steve
>

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v

Enterprise-wide editing can require a lot of infrastructure to support it.
A large-scale
implementation might include (this is only one scenario):

 o  a data repository  / warehouse / database
 o  project workspaces for editing
 o  a "view-only" copy of the data

Typical workflows would include:

1. Edit: operator identifies features in warehouse for editing, locks them,
extracts them
    to the project workspace.  The features are edited, possibly reviewed,
then checked
    back into the warehouse.  This is sometimes known as a "long
transaction"
    Some things that may be important:
        1. feature level locking (as apposed to file locking) to prevent
simultaneous editing
        2. feature lineage tracking: timestamps, feature "retirement"
instead of deletion
        3. theme security: certain departments can edit only specific themes

2. Copy:  at a pre-determined schedule, the warehouse is copied to the
"View-only"
    database.  This may include re-formatting, indexing and distributing the
data to get better
    performance for viewing.  Depending on the edits, the copy could be once
a day,
    once a month, etc.  The good thing about this approach is that the user
    (viewer/querier) has a stable data set to operate on.  The bad thing is
it might not be
    up to date.

3. Viewing: the data is queried and rendered for thick and thin client apps.

Of course all this might be unnecessary if you only have occasional edits
and a few
viewers....

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v


> I'm glad to hear I'm not the only one having problems with updating
> shapefiles. :)
>
> >From looking at the shapefile definition paper, you can see that there is
an
> index file .SHX which points to a .SHP file which has variable length
records.
>
> There are a few problems that I can see.  Please verify if any of these
are
> correct or not.
> 1. Deleting an old object.  I think this can be handled by setting the
> shapetype to a NULL shape.
>
> 2. Increasing the number of vertices of a shape, and hence increasing the
> record size.  I think the best way to handle this is to remove the old
shape
> by setting its shapetype to NULL, and to add a new shape to the end of the
> .SHP file.  The pointer in the .SHX file will now have to be redirected to
the
> end of the .SHP file.  This now means that the order of the .SHP file and
the
> .SHX file will not match, which will reduce query speeds, so periodically
the
> datafiles would need to be rebuilt.
>
> 3. There is an issue with the .SHP file and .SHX file becoming out of
sync.
> Basically, when a shape is updated, first the .SHP file will need to be
> updated, and some time later the .SHX file will be updated.  There is a
window
> of opportunity where the files will be out of sync.  I was planning to
address
> this by either putting in a lock file, or changing read/write permissions
to
> lock the files while the database is out of sync.
> This means that some reads of the database will fail because the database
is
> updating.
>
> 4. I'm not sure what the best way is to link into a SQL database.  If the
> shapefile is only added to, then the best way to reference an object is by
> using the index in the .SHX file.  However, if you delete an object,
should
> you rebuild the .SHX file?  This will keep the index file from blowing
out,
> but all the indexes will change and hence the  SQL database will reference
the
> wrong indices.

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v
How about a unique key stored in the dbf file used to join to the SQL
database?

This would allow for many shapefiles joining to a single SQL table (might be
useful if the data is tiled.)
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v


>
> Happy for any advice.
>
> Cameron.
>



From steve.lime at dnr.state.mn.us  Tue Oct 12 14:36:38 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 12 Oct 1999 16:36:38 -0500
Subject: Updating Shapefiles, and data integrity
Message-ID: <s803639e.015@smtp.dnr.state.mn.us>

The more I think about this, the more I understand why I never attempted it.
Locking is a real pain in the CGI world. When do you lock, when a record
is requested or when edits are submitted? If the latter then there is a chance
more than one person could request the same shape. I don't think that 
on-the-fly edits are possible robustly. Somehow I think edits need to be cached
and commited "behind the scenes". It's essential the shp, shx and dbf records
remain in sync. What about something like this:

Assuming the there is a mechanism to request a shape and attributes (and a checkout
time)  and make changes. A user now sends back some edits. This causes a record to 
be written to a "pending" database. What actually gets saved are things like source 
shapefile, feature id (-1 for new), timestamp, etc. The actually edits get saved in some format 
(shapefile) as a  file whose name can be reconstructed from elements in the pending database. 
Now, periodically a process could go through and commit the edits in the pending database
to production (not web accessible) versions. When this is finished the updated stuff
could be swapped in for the old stuff and the pending database purged (in part). The 
commiting of the shapes would essentially involve rebuilding the shapefile from the user
edits and the production version (i.e. pick the edited version if it exists). Put a lock in
place while versions are being swapped and remove it when done, probably only a
few seconds. You could even maintain a history by saving previous versions for
some period of time or retiring shapes to some external format (shapefile).

As requests for shapes come in a quick check of the pending database could be used
to identify re-edits. If a timestamp is set when a shape is requested then it could be
compared against edits in the pending database to identify possible problems. If a user
requests an edited shape just send the pending edits as if they were part of the current
shapefile. New shapes are just added to the pending database and make their way
into the main database as part of the update process.

Sounds complicated but really is only 2 processes, 1 database and a bunch of cached
edits. Timestamps can help alleviate simultaneous edits and a the worst thing a user
would see would be a message like "The record you're submitting has changed since
you requested it, cannot process the edit. Would you like to work from the edited version?".

Again, without some sort of a persistant connection I doubt that real-time editing is possible.
One could bump the commit time up to a few minutes or even seconds though so it would
certainly seem real-time.

(Cameron, this approach would involve no editing of Frank's shapelib at all since all you're
doing is reading and writing individual records. The effort goes into getting all the communication
working right. Could even be a perl script with system calls to the shapelib utils for creating
files and adding info.)

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

>>> "bfraser" <bfraser at geoanalytic.ab.ca> 10/12 11:00 AM >>>
see my comments below...

Brent Fraser


----- Original Message -----
From: Cameron Shorter <cshorter at optusnet.com.au>
To: mapserver <mapserver-users at lists.gis.umn.edu>
Sent: Sunday, October 10, 1999 4:14 AM
Subject: Updating Shapefiles, and data integrity


>
>
> -------- Original Message --------
> Subject: Re: request for comments...
> Date: Sun, 10 Oct 1999 19:16:24 +1000
> From: Cameron Shorter <cshorter at optusnet.com.au>
> Reply-To: camerons at cat.org.au 
> To: Stephen Lime <steve.lime at dnr.state.mn.us>
> References: <s7ff017e.048 at smtp.dnr.state.mn.us>
>
>
>
> Stephen Lime wrote:
> >
> > Maintaining data integrity is going to be a big issue. I was at our
state GIS conference and got to chat with Jack Dangermond from ESRI about
the MapServer and their new ArcIMS product. Seems they're having trouble
with this editing stuff. Shapefiles just aren't a transactional environment
so unless you can assure yourself of single user access there's always the
potential for multiple concurrent edits. Then there's the issue of quality
control. I think the solution needs to offer immediate update and delayed
update. ArcIMS, as I understand it, caches updates until an operator on the
server site commits the edits to the main database. This operator could be a
cron process I suppose that could handle locking while
> > edits are processed. I think this may be a good approach as you could do
some simple transaction management- review, edit and delete, once the
initial work was done. Edits could be stored in a shapefile along with
attributes and enough additional information to commit the shape - source
shapefile, shape (or is a new one), type of edit (replace, attribute change)
etc.
> >
> > Anyway, just my thoughts...
> >
> > Steve
>

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v

Enterprise-wide editing can require a lot of infrastructure to support it.
A large-scale
implementation might include (this is only one scenario):

 o  a data repository  / warehouse / database
 o  project workspaces for editing
 o  a "view-only" copy of the data

Typical workflows would include:

1. Edit: operator identifies features in warehouse for editing, locks them,
extracts them
    to the project workspace.  The features are edited, possibly reviewed,
then checked
    back into the warehouse.  This is sometimes known as a "long
transaction"
    Some things that may be important:
        1. feature level locking (as apposed to file locking) to prevent
simultaneous editing
        2. feature lineage tracking: timestamps, feature "retirement"
instead of deletion
        3. theme security: certain departments can edit only specific themes

2. Copy:  at a pre-determined schedule, the warehouse is copied to the
"View-only"
    database.  This may include re-formatting, indexing and distributing the
data to get better
    performance for viewing.  Depending on the edits, the copy could be once
a day,
    once a month, etc.  The good thing about this approach is that the user
    (viewer/querier) has a stable data set to operate on.  The bad thing is
it might not be
    up to date.

3. Viewing: the data is queried and rendered for thick and thin client apps.

Of course all this might be unnecessary if you only have occasional edits
and a few
viewers....

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v


> I'm glad to hear I'm not the only one having problems with updating
> shapefiles. :)
>
> >From looking at the shapefile definition paper, you can see that there is
an
> index file .SHX which points to a .SHP file which has variable length
records.
>
> There are a few problems that I can see.  Please verify if any of these
are
> correct or not.
> 1. Deleting an old object.  I think this can be handled by setting the
> shapetype to a NULL shape.
>
> 2. Increasing the number of vertices of a shape, and hence increasing the
> record size.  I think the best way to handle this is to remove the old
shape
> by setting its shapetype to NULL, and to add a new shape to the end of the
> .SHP file.  The pointer in the .SHX file will now have to be redirected to
the
> end of the .SHP file.  This now means that the order of the .SHP file and
the
> .SHX file will not match, which will reduce query speeds, so periodically
the
> datafiles would need to be rebuilt.
>
> 3. There is an issue with the .SHP file and .SHX file becoming out of
sync.
> Basically, when a shape is updated, first the .SHP file will need to be
> updated, and some time later the .SHX file will be updated.  There is a
window
> of opportunity where the files will be out of sync.  I was planning to
address
> this by either putting in a lock file, or changing read/write permissions
to
> lock the files while the database is out of sync.
> This means that some reads of the database will fail because the database
is
> updating.
>
> 4. I'm not sure what the best way is to link into a SQL database.  If the
> shapefile is only added to, then the best way to reference an object is by
> using the index in the .SHX file.  However, if you delete an object,
should
> you rebuild the .SHX file?  This will keep the index file from blowing
out,
> but all the indexes will change and hence the  SQL database will reference
the
> wrong indices.

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v
How about a unique key stored in the dbf file used to join to the SQL
database?

This would allow for many shapefiles joining to a single SQL table (might be
useful if the data is tiled.)
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
v


>
> Happy for any advice.
>
> Cameron.
>



From SullivanJ at nima.mil  Wed Oct 13 05:51:32 1999
From: SullivanJ at nima.mil (Sullivan, James R.)
Date: Wed, 13 Oct 1999 08:51:32 -0400
Subject: Updating Shapefiles, and data integrity
Message-ID: <29E2BCE114FBD1118EE000805FBEA41380B8F1@WNYX2>

Believe a better approach would be to develop a Mapserver interface to
Oracle Spatial or ESRI SDE.  Let the database do all the locking, etc.
This would answer the mail on a number of other issues, too.  Both have open
C api that can be downloaded via the net.


Jim Sullivan
NIMA / TES

	-----Original Message-----
	From:	Stephen Lime [SMTP:steve.lime at dnr.state.mn.us]
	Sent:	Tuesday, October 12, 1999 5:37 PM
	To:	camerons at cat.org.au; bfraser at geoanalytic.ab.ca;
mapserver-users at lists.gis.umn.edu
	Subject:	Re: Updating Shapefiles, and data integrity

	The more I think about this, the more I understand why I never
attempted it.
	Locking is a real pain in the CGI world. When do you lock, when a
record
	is requested or when edits are submitted? If the latter then there
is a chance
	more than one person could request the same shape. I don't think
that 
	on-the-fly edits are possible robustly. Somehow I think edits need
to be cached
	and commited "behind the scenes". It's essential the shp, shx and
dbf records
	remain in sync. What about something like this:

	Assuming the there is a mechanism to request a shape and attributes
(and a checkout
	time)  and make changes. A user now sends back some edits. This
causes a record to 
	be written to a "pending" database. What actually gets saved are
things like source 
	shapefile, feature id (-1 for new), timestamp, etc. The actually
edits get saved in some format 
	(shapefile) as a  file whose name can be reconstructed from elements
in the pending database. 
	Now, periodically a process could go through and commit the edits in
the pending database
	to production (not web accessible) versions. When this is finished
the updated stuff
	could be swapped in for the old stuff and the pending database
purged (in part). The 
	commiting of the shapes would essentially involve rebuilding the
shapefile from the user
	edits and the production version (i.e. pick the edited version if it
exists). Put a lock in
	place while versions are being swapped and remove it when done,
probably only a
	few seconds. You could even maintain a history by saving previous
versions for
	some period of time or retiring shapes to some external format
(shapefile).

	As requests for shapes come in a quick check of the pending database
could be used
	to identify re-edits. If a timestamp is set when a shape is
requested then it could be
	compared against edits in the pending database to identify possible
problems. If a user
	requests an edited shape just send the pending edits as if they were
part of the current
	shapefile. New shapes are just added to the pending database and
make their way
	into the main database as part of the update process.

	Sounds complicated but really is only 2 processes, 1 database and a
bunch of cached
	edits. Timestamps can help alleviate simultaneous edits and a the
worst thing a user
	would see would be a message like "The record you're submitting has
changed since
	you requested it, cannot process the edit. Would you like to work
from the edited version?".

	Again, without some sort of a persistant connection I doubt that
real-time editing is possible.
	One could bump the commit time up to a few minutes or even seconds
though so it would
	certainly seem real-time.

	(Cameron, this approach would involve no editing of Frank's shapelib
at all since all you're
	doing is reading and writing individual records. The effort goes
into getting all the communication
	working right. Could even be a perl script with system calls to the
shapelib utils for creating
	files and adding info.)

	Steve

	Stephen Lime
	Internet Applications Analyst
	MIS Bureau - MN DNR

	(651) 297-2937
	steve.lime at dnr.state.mn.us

	>>> "bfraser" <bfraser at geoanalytic.ab.ca> 10/12 11:00 AM >>>
	see my comments below...

	Brent Fraser


	----- Original Message -----
	From: Cameron Shorter <cshorter at optusnet.com.au>
	To: mapserver <mapserver-users at lists.gis.umn.edu>
	Sent: Sunday, October 10, 1999 4:14 AM
	Subject: Updating Shapefiles, and data integrity


	>
	>
	> -------- Original Message --------
	> Subject: Re: request for comments...
	> Date: Sun, 10 Oct 1999 19:16:24 +1000
	> From: Cameron Shorter <cshorter at optusnet.com.au>
	> Reply-To: camerons at cat.org.au 
	> To: Stephen Lime <steve.lime at dnr.state.mn.us>
	> References: <s7ff017e.048 at smtp.dnr.state.mn.us>
	>
	>
	>
	> Stephen Lime wrote:
	> >
	> > Maintaining data integrity is going to be a big issue. I was at
our
	state GIS conference and got to chat with Jack Dangermond from ESRI
about
	the MapServer and their new ArcIMS product. Seems they're having
trouble
	with this editing stuff. Shapefiles just aren't a transactional
environment
	so unless you can assure yourself of single user access there's
always the
	potential for multiple concurrent edits. Then there's the issue of
quality
	control. I think the solution needs to offer immediate update and
delayed
	update. ArcIMS, as I understand it, caches updates until an operator
on the
	server site commits the edits to the main database. This operator
could be a
	cron process I suppose that could handle locking while
	> > edits are processed. I think this may be a good approach as you
could do
	some simple transaction management- review, edit and delete, once
the
	initial work was done. Edits could be stored in a shapefile along
with
	attributes and enough additional information to commit the shape -
source
	shapefile, shape (or is a new one), type of edit (replace, attribute
change)
	etc.
	> >
	> > Anyway, just my thoughts...
	> >
	> > Steve
	>

	
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
	v

	Enterprise-wide editing can require a lot of infrastructure to
support it.
	A large-scale
	implementation might include (this is only one scenario):

	 o  a data repository  / warehouse / database
	 o  project workspaces for editing
	 o  a "view-only" copy of the data

	Typical workflows would include:

	1. Edit: operator identifies features in warehouse for editing,
locks them,
	extracts them
	    to the project workspace.  The features are edited, possibly
reviewed,
	then checked
	    back into the warehouse.  This is sometimes known as a "long
	transaction"
	    Some things that may be important:
	        1. feature level locking (as apposed to file locking) to
prevent
	simultaneous editing
	        2. feature lineage tracking: timestamps, feature
"retirement"
	instead of deletion
	        3. theme security: certain departments can edit only
specific themes

	2. Copy:  at a pre-determined schedule, the warehouse is copied to
the
	"View-only"
	    database.  This may include re-formatting, indexing and
distributing the
	data to get better
	    performance for viewing.  Depending on the edits, the copy could
be once
	a day,
	    once a month, etc.  The good thing about this approach is that
the user
	    (viewer/querier) has a stable data set to operate on.  The bad
thing is
	it might not be
	    up to date.

	3. Viewing: the data is queried and rendered for thick and thin
client apps.

	Of course all this might be unnecessary if you only have occasional
edits
	and a few
	viewers....

	
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
	v


	> I'm glad to hear I'm not the only one having problems with
updating
	> shapefiles. :)
	>
	> >From looking at the shapefile definition paper, you can see that
there is
	an
	> index file .SHX which points to a .SHP file which has variable
length
	records.
	>
	> There are a few problems that I can see.  Please verify if any of
these
	are
	> correct or not.
	> 1. Deleting an old object.  I think this can be handled by setting
the
	> shapetype to a NULL shape.
	>
	> 2. Increasing the number of vertices of a shape, and hence
increasing the
	> record size.  I think the best way to handle this is to remove the
old
	shape
	> by setting its shapetype to NULL, and to add a new shape to the
end of the
	> .SHP file.  The pointer in the .SHX file will now have to be
redirected to
	the
	> end of the .SHP file.  This now means that the order of the .SHP
file and
	the
	> .SHX file will not match, which will reduce query speeds, so
periodically
	the
	> datafiles would need to be rebuilt.
	>
	> 3. There is an issue with the .SHP file and .SHX file becoming out
of
	sync.
	> Basically, when a shape is updated, first the .SHP file will need
to be
	> updated, and some time later the .SHX file will be updated.  There
is a
	window
	> of opportunity where the files will be out of sync.  I was
planning to
	address
	> this by either putting in a lock file, or changing read/write
permissions
	to
	> lock the files while the database is out of sync.
	> This means that some reads of the database will fail because the
database
	is
	> updating.
	>
	> 4. I'm not sure what the best way is to link into a SQL database.
If the
	> shapefile is only added to, then the best way to reference an
object is by
	> using the index in the .SHX file.  However, if you delete an
object,
	should
	> you rebuild the .SHX file?  This will keep the index file from
blowing
	out,
	> but all the indexes will change and hence the  SQL database will
reference
	the
	> wrong indices.

	
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
	v
	How about a unique key stored in the dbf file used to join to the
SQL
	database?

	This would allow for many shapefiles joining to a single SQL table
(might be
	useful if the data is tiled.)
	
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
	v


	>
	> Happy for any advice.
	>
	> Cameron.
	>


From david.fawcett at moea.state.mn.us  Wed Oct 13 06:19:18 1999
From: david.fawcett at moea.state.mn.us (Fawcett, David)
Date: Wed, 13 Oct 1999 08:19:18 -0500
Subject: FW: Updating Shapefiles, and data integrity
Message-ID: <3E2112BBDA46D311824C00805F9FE4FF40B0DE@pca-ex03.pca.state.mn.us>

This may be fine as an "extension", but if it became integral, it would certainly change the product from a free program to a pretty expensive one.

Even if a non-commercial solution to this issue is a little kludgy and slower than real-time, it may serve most people's needs, and will avoid spendy Oracle or SDE licenses.

David Fawcett
Minnesota Office of Environmental Assistance

> ----------
> From: 	Sullivan, James R.[SMTP:SullivanJ at nima.mil]
> Sent: 	Wednesday, October 13, 1999 7:51 AM
> To: 	'Stephen Lime'; camerons at cat.org.au; bfraser at geoanalytic.ab.ca; mapserver-users at lists.gis.umn.edu
> Subject: 	RE: Updating Shapefiles, and data integrity
> 
> Believe a better approach would be to develop a Mapserver interface to
> Oracle Spatial or ESRI SDE.  Let the database do all the locking, etc.
> This would answer the mail on a number of other issues, too.  Both have open
> C api that can be downloaded via the net.
> 
> 
> Jim Sullivan
> NIMA / TES
> 
> 	-----Original Message-----
> 	From:	Stephen Lime [SMTP:steve.lime at dnr.state.mn.us]
> 	Sent:	Tuesday, October 12, 1999 5:37 PM
> 	To:	camerons at cat.org.au; bfraser at geoanalytic.ab.ca;
> mapserver-users at lists.gis.umn.edu
> 	Subject:	Re: Updating Shapefiles, and data integrity
> 
> 	The more I think about this, the more I understand why I never
> attempted it.
> 	Locking is a real pain in the CGI world. When do you lock, when a
> record
> 	is requested or when edits are submitted? If the latter then there
> is a chance
> 	more than one person could request the same shape. I don't think
> that 
> 	on-the-fly edits are possible robustly. Somehow I think edits need
> to be cached
> 	and commited "behind the scenes". It's essential the shp, shx and
> dbf records
> 	remain in sync. What about something like this:
> 
> 	Assuming the there is a mechanism to request a shape and attributes
> (and a checkout
> 	time)  and make changes. A user now sends back some edits. This
> causes a record to 
> 	be written to a "pending" database. What actually gets saved are
> things like source 
> 	shapefile, feature id (-1 for new), timestamp, etc. The actually
> edits get saved in some format 
> 	(shapefile) as a  file whose name can be reconstructed from elements
> in the pending database. 
> 	Now, periodically a process could go through and commit the edits in
> the pending database
> 	to production (not web accessible) versions. When this is finished
> the updated stuff
> 	could be swapped in for the old stuff and the pending database
> purged (in part). The 
> 	commiting of the shapes would essentially involve rebuilding the
> shapefile from the user
> 	edits and the production version (i.e. pick the edited version if it
> exists). Put a lock in
> 	place while versions are being swapped and remove it when done,
> probably only a
> 	few seconds. You could even maintain a history by saving previous
> versions for
> 	some period of time or retiring shapes to some external format
> (shapefile).
> 
> 	As requests for shapes come in a quick check of the pending database
> could be used
> 	to identify re-edits. If a timestamp is set when a shape is
> requested then it could be
> 	compared against edits in the pending database to identify possible
> problems. If a user
> 	requests an edited shape just send the pending edits as if they were
> part of the current
> 	shapefile. New shapes are just added to the pending database and
> make their way
> 	into the main database as part of the update process.
> 
> 	Sounds complicated but really is only 2 processes, 1 database and a
> bunch of cached
> 	edits. Timestamps can help alleviate simultaneous edits and a the
> worst thing a user
> 	would see would be a message like "The record you're submitting has
> changed since
> 	you requested it, cannot process the edit. Would you like to work
> from the edited version?".
> 
> 	Again, without some sort of a persistant connection I doubt that> 
> real-time editing is possible.
> 	One could bump the commit time up to a few minutes or even seconds
> though so it would
> 	certainly seem real-time.
> 
> 	(Cameron, this approach would involve no editing of Frank's shapelib
> at all since all you're
> 	doing is reading and writing individual records. The effort goes
> into getting all the communication
> 	working right. Could even be a perl script with system calls to the
> shapelib utils for creating
> 	files and adding info.)
> 
> 	Steve
> 
> 	Stephen Lime
> 	Internet Applications Analyst
> 	MIS Bureau - MN DNR
> 
> 	(651) 297-2937
> 	steve.lime at dnr.state.mn.us
> 
> 	>>> "bfraser" <bfraser at geoanalytic.ab.ca> 10/12 11:00 AM >>>
> 	see my comments below...
> 
> 	Brent Fraser
> 
> 
> 	----- Original Message -----
> 	From: Cameron Shorter <cshorter at optusnet.com.au>
> 	To: mapserver <mapserver-users at lists.gis.umn.edu>
> 	Sent: Sunday, October 10, 1999 4:14 AM
> 	Subject: Updating Shapefiles, and data integrity
> 
> 
> 	>
> 	>
> 	> -------- Original Message --------
> 	> Subject: Re: request for comments...
> 	> Date: Sun, 10 Oct 1999 19:16:24 +1000
> 	> From: Cameron Shorter <cshorter at optusnet.com.au>
> 	> Reply-To: camerons at cat.org.au 
> 	> To: Stephen Lime <steve.lime at dnr.state.mn.us>
> 	> References: <s7ff017e.048 at smtp.dnr.state.mn.us>
> 	>
> 	>
> 	>
> 	> Stephen Lime wrote:
> 	> >
> 	> > Maintaining data integrity is going to be a big issue. I was at
> our
> 	state GIS conference and got to chat with Jack Dangermond from ESRI
> about
> 	the MapServer and their new ArcIMS product. Seems they're having
> trouble
> 	with this editing stuff. Shapefiles just aren't a transactional
> environment
> 	so unless you can assure yourself of single user access there's
> always the
> 	potential for multiple concurrent edits. Then there's the issue of
> quality
> 	control. I think the solution needs to offer immediate update and
> delayed
> 	update. ArcIMS, as I understand it, caches updates until an operator
> on the
> 	server site commits the edits to the main database. This operator
> could be a
> 	cron process I suppose that could handle locking while
> 	> > edits are processed. I think this may be a good approach as you
> could do
> 	some simple transaction management- review, edit and delete, once
> the
> 	initial work was done. Edits could be stored in a shapefile along
> with
> 	attributes and enough additional information to commit the shape -
> source
> 	shapefile, shape (or is a new one), type of edit (replace, attribute
> change)
> 	etc.
> 	> >
> 	> > Anyway, just my thoughts...
> 	> >
> 	> > Steve
> 	>
> 
> 	
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> 	v
> 
> 	Enterprise-wide editing can require a lot of infrastructure to
> support it.
> 	A large-scale
> 	implementation might include (this is only one scenario):
> 
> 	 o  a data repository  / warehouse / database
> 	 o  project workspaces for editing
> 	 o  a "view-only" copy of the data
> 
> 	Typical workflows would include:
> 
> 	1. Edit: operator identifies features in warehouse for editing,
> locks them,
> 	extracts them
> 	    to the project workspace.  The features are edited, possibly
> reviewed,
> 	then checked
> 	    back into the warehouse.  This is sometimes known as a "long
> 	transaction"
> 	    Some things that may be important:
> 	        1. feature level locking (as apposed to file locking) to
> prevent
> 	simultaneous editing
> 	        2. feature lineage tracking: timestamps, feature
> "retirement"
> 	instead of deletion
> 	        3. theme security: certain departments can edit only
> specific themes
> 
> 	2. Copy:  at a pre-determined schedule, the warehouse is copied to
> the
> 	"View-only"
> 	    database.  This may include re-formatting, indexing and
> distributing the
> 	data to get better
> 	    performance for viewing.  Depending on the edits, the copy could
> be once
> 	a day,
> 	    once a month, etc.  The good thing about this approach is that> 
> the user
> 	    (viewer/querier) has a stable data set to operate on.  The bad
> thing is
> 	it might not be
> 	    up to date.
> 
> 	3. Viewing: the data is queried and rendered for thick and thin
> client apps.
> 
> 	Of course all this might be unnecessary if you only have occasional
> edits
> 	and a few
> 	viewers....
> 
> 	
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> 	v
> 
> 
> 	> I'm glad to hear I'm not the only one having problems with
> updating
> 	> shapefiles. :)
> 	>
> 	> >From looking at the shapefile definition paper, you can see that
> there is
> 	an
> 	> index file .SHX which points to a .SHP file which has variable
> length
> 	records.
> 	>
> 	> There are a few problems that I can see.  Please verify if any of
> these
> 	are
> 	> correct or not.
> 	> 1. Deleting an old object.  I think this can be handled by setting
> the
> 	> shapetype to a NULL shape.
> 	>
> 	> 2. Increasing the number of vertices of a shape, and hence
> increasing the
> 	> record size.  I think the best way to handle this is to remove the
> old
> 	shape
> 	> by setting its shapetype to NULL, and to add a new shape to the
> end of the
> 	> .SHP file.  The pointer in the .SHX file will now have to be
> redirected to
> 	the
> 	> end of the .SHP file.  This now means that the order of the .SHP
> file and
> 	the
> 	> .SHX file will not match, which will reduce query speeds, so
> periodically
> 	the
> 	> datafiles would need to be rebuilt.
> 	>
> 	> 3. There is an issue with the .SHP file and .SHX file becoming out
> of
> 	sync.
> 	> Basically, when a shape is updated, first the .SHP file will need
> to be
> 	> updated, and some time later the .SHX file will be updated.  There
> is a
> 	window
> 	> of opportunity where the files will be out of sync.  I was
> planning to
> 	address
> 	> this by either putting in a lock file, or changing read/write
> permissions
> 	to
> 	> lock the files while the database is out of sync.
> 	> This means that some reads of the database will fail because the
> database
> 	is
> 	> updating.
> 	>
> 	> 4. I'm not sure what the best way is to link into a SQL database.
> If the
> 	> shapefile is only added to, then the best way to reference an
> object is by
> 	> using the index in the .SHX file.  However, if you delete an
> object,
> 	should
> 	> you rebuild the .SHX file?  This will keep the index file from
> blowing
> 	out,
> 	> but all the indexes will change and hence the  SQL database will
> reference
> 	the
> 	> wrong indices.
> 
> 	
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> 	v
> 	How about a unique key stored in the dbf file used to join to the
> SQL
> 	database?
> 
> 	This would allow for many shapefiles joining to a single SQL table
> (might be
> 	useful if the data is tiled.)
> 	
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> 	v
> 
> 
> 	>
> 	> Happy for any advice.
> 	>
> 	> Cameron.
> 	>
> 


From pgallen at randomlogic.com  Wed Oct 13 13:09:18 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Wed, 13 Oct 1999 13:09:18 -0700
Subject: FW: Updating Shapefiles, and data integrity
References: <3E2112BBDA46D311824C00805F9FE4FF40B0DE@pca-ex03.pca.state.mn.us>
Message-ID: <3804E6EE.2D7CC928@randomlogic.com>

"Fawcett, David" wrote:

> This may be fine as an "extension", but if it became integral, it would certainly change the product from a free program to a pretty expensive one.
>
> Even if a non-commercial solution to this issue is a little kludgy and slower than real-time, it may serve most people's needs, and will avoid spendy Oracle or SDE licenses.
>

[BIG_HUGE_SNIP]

I know little about database software, but wouldn't something like Postgres SQL  (postgresql) or MySQL work for this? Also, Oracle 8i and IBM's DB2 are free for personal use and
development under Linux. Would either of those work? My brother (not on this list, but maybe he should be) said it sounds to him like what is needed is a relational database to
handle the three files - .dbf, .shp, .shx. He also said the only database software he knew of that handles record locking is Paradox, where the others use file locking.

PGA



From pgallen at randomlogic.com  Wed Oct 13 13:35:03 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Wed, 13 Oct 1999 13:35:03 -0700
Subject: Making .dbf, .shp, and .shx files
Message-ID: <3804ECF7.4E6DB9A6@randomlogic.com>

At the risk of looking stupid...

A couple people told me I could make shape files containing the shapes of all
the county maps I have so that I can tile them into one larger map. This makes
perfect sense to me. Since every county map already has a county shape file, I
already have the shape data. Since I am quite a novice at this: How do I combine
all the various shape files into a single shape file? How do I combine the
various .shx and .dbf that accompany these shape files into corresponding files
to go along with the single county shape file?

What would really be nice is:

1. Have a map of the entire continental US with each state outlined, and a
couple smaller maps containing the US territories (American Samoa, the Pacific
Territories, Alaska, etc.).
2. The user clicks on a state/territory and is then given another map of that
state with the various counties outlines.
3. The user can then zoom/pan/query that map to find what he/she needs.

I've read in to documentation that it's possible to set a maximum scale so that
the interface can change from one map to another - zoom out too far in a county
map, and the entire state is displayed, or maybe the surrounding counties are
tiled to make a slightly larger (more coverage)  map. I assume you can get the
same result in the other direction? (Zoom in too far in the state map and you
get a county or several counties.) What if you pan off the edge of a map?
Something else I'd like to do in the immediate future: Plot my own data on a
map. The data is retrieved via text email messages. The data includes LAT/LONG
position and various pieces of information regarding the status of a field unit.
I'd like to be able to take this type of data from an arbitrary number of units,
concatenate it into a single set of ESRI files, and display/query it just like
I'm doing with current map data. How difficult is this and are the tools for it
already included/supplied? (remember, I'm a novice at this and only have about
16hrs. total experience with MapServer - the first 8 involved making the first
map of San Diego :)

BTW, I've updated my San Diego map and added Cook County, IL to my test web site
with enhanced 1997-98 data from the Census. (My wife used to live near Chicago
:) Again the URL is:
24.5.14.144 or http://cx570538-a.elcjn1.sdca.home.com.

PGA





From mjacquin at geosys-inc.com  Wed Oct 13 16:37:26 1999
From: mjacquin at geosys-inc.com (Marc JACQUIN - GEOSYS Inc - Minneapolis)
Date: Wed, 13 Oct 1999 16:37:26 -0700
Subject: Data integrity in shape/dbf files
Message-ID: <380517B6.7422F1A1@geosys-inc.com>

Hi everyone,

Seems to be a great deal to be able to have concurrent updates on shape
files (and dbf).

In an application only using point shapes, I tried to implement record
locking on DBF. I am only using MapServer functions calls to generate
GIF files (from a map file) and embed it into my own CGI. Seems to work
on Win32 with lock functions. On Unix there is an equivalent through
fnctl. I added those functions to the shapelib. Of course, it is only
valid for attributes. Must be careful about dead-locking.

Due to Web browsing, we cannot assure that the end user has an updated
view of the database. There is no way with CGI to know if a page needs
to be reloaded, but when an update is done the CGI can reload the main
page after the post, so the end user sees the changes he did.

Changing shapes cannot really be done with the ismap capabilities. Using
some Java is required. My idea is that shape files and objects can be
stored into a DBMS as BLOB (like SDE does I guess).

Marc
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mjacquin.vcf
Type: text/x-vcard
Size: 416 bytes
Desc: Card for Marc JACQUIN - GEOSYS Inc - Minneapolis 
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991013/10f6a09b/attachment.vcf>

From pgallen at randomlogic.com  Wed Oct 13 14:46:00 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Wed, 13 Oct 1999 14:46:00 -0700
Subject: Colr look up table
Message-ID: <3804FD98.C9EFAEB1@randomlogic.com>

On the ForNet site I was reading some old MapServer documentation (v2.3) about
making a color table. Does the latest version of MapServer support this? I have
a file as outlined in the afore mentioned documentation and I have used the old
printcolors utility to create a GIF image of it. Now I'm wondering if/how I can
use the color file.

Thanks,

PGA



From Ross.Searle at dnr.qld.gov.au  Thu Oct 14 00:16:41 1999
From: Ross.Searle at dnr.qld.gov.au (Searle Ross)
Date: Thu, 14 Oct 1999 17:16:41 +1000
Subject: No subject
Message-ID: <C10DC4222C68D311B7E50008C79F8B12042DB2@enterprise.dnr.qld.gov.au>

Hi

Anyone got any ideas about what might be causing Version 3.22 to return the
following error from the web server

[Thu Oct 14 16:53:23 1999] HTTPd: malformed header from script
/usr/local/http/cgi-bin/mapserv3.22
[Thu Oct 14 16:53:23 1999] HTTPd: send aborted for 161.143.17.153, URL:
/cgi-bin/mapserv3.22

I have version 3.1 working fine.


my ini file is as follows



# Where are fonts and symbols files located
SYMBOL_PATH=/data4/WWW/mapserver/symbols

# Absolute path to writable directory where images are to be written. This
# directory needs to be:
#   (1) writable by the user httpd is running as 
#   (2) accessible via the web
IMAGE_PATH=/data4/WWW/tmp

# Path relative to document root of the web server for the writeable
directory.
IMAGE_VIRTUAL_PATH=tmp

# Contact information provided if errors occur.
CONTACT_EMAIL=ross.searle at dnr.qld.gov.au

# Where are the log files to be created. Comment out if you don't want them.
# You must create the directory and make it writable by the user the web
server
# runs as. Alternatively, create a file called "access.log" there and change
# ownership or permissions accordingly.
LOG_PATH=/data4/WWW/mapserver

# File extention of template html files. Should be a regular expression.
This
# usually doesn't need to be changed.
#TEMPLATE_EXTENSION=\.html

# File extention of map files. Should be a regular expression. This usually
# doesn't need to be changed.
#MAP_EXTENSION=\.map





Ross Searle
Natural Resource Information Management Officer
Department of Natural Resources
PO Box 1143 
Bundaberg QLD 4670. Australia
Ph : 0741 537 888 Fax : 0741 537 823



From pgallen at randomlogic.com  Thu Oct 14 00:53:37 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Thu, 14 Oct 1999 00:53:37 -0700
Subject: No subject
References: <C10DC4222C68D311B7E50008C79F8B12042DB2@enterprise.dnr.qld.gov.au>
Message-ID: <38058C01.441EDFD3@randomlogic.com>

Searle Ross wrote:

> Hi
>
> Anyone got any ideas about what might be causing Version 3.22 to return the
> following error from the web server
>
> [Thu Oct 14 16:53:23 1999] HTTPd: malformed header from script
> /usr/local/http/cgi-bin/mapserv3.22
> [Thu Oct 14 16:53:23 1999] HTTPd: send aborted for 161.143.17.153, URL:
> /cgi-bin/mapserv3.22
>
> I have version 3.1 working fine.
>

3.22? Didn't know this was out. I checked the site tonight and saw the same 3.21
10/01/99 listed. Guess I'll have to update (once there's a clear answer to your
problem. I'm using 3.20 now)

PGA





From pgallen at randomlogic.com  Thu Oct 14 05:14:32 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Thu, 14 Oct 1999 05:14:32 -0700
Subject: Once again I've partially answered my own questions
Message-ID: <3805C928.4DB57756@randomlogic.com>

By reading a little about the Shapelib API and writing a small application, I
have been able to concatenate shapefile data painlessly. I also know how I can
convert the data from our field units into a format that can be used by
MapServer. I'd still like to know if/how a color table can be used though.

The only missing piece now seems to be a decent TSP (Travelling Salesman
Problem) program and vehicle routing program. So far what I've seen on the web
is not really what we need: The VRP data I've seen deals with forms of the TSP
with Time Windows, when all I want is an algorithm to tell me how to get from
point A to point B within a city, not from city to city. So far the quick and
easy solution eludes me, but if I have to write my own, so be it. :)

Thanks to all those who've given me insight into the workings of MapServer,
shapefiles, and Shapelib.

PGA


From mapadm at polsl.gliwice.pl  Thu Oct 14 05:45:53 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Thu, 14 Oct 1999 14:45:53 +0200
Subject: Once again I've partially answered my own questions
Message-ID: <004101bf1642$114c7820$0900000a@pc-9.ispik.pl>

-----Original Message-----
From: Paul G. Allen <pgallen at randomlogic.com>
To: MapServer Users List <mapserver-users at lists.gis.umn.edu>
Date: 14 pa?dziernika 1999 14:16
Subject: Once again I've partially answered my own questions


>MapServer. I'd still like to know if/how a color table can be used though.


There is no color table in current version of MapServer. Instead you use RGB
values directly.

Grzegorz



From mjacquin at geosys-inc.com  Thu Oct 14 10:05:19 1999
From: mjacquin at geosys-inc.com (Marc JACQUIN - GEOSYS Inc - Minneapolis)
Date: Thu, 14 Oct 1999 10:05:19 -0700
Subject: [MAPSERVER] Message headers
Message-ID: <38060D4E.1A4F8C8F@geosys-inc.com>

Could we have a common tag inside message headers so we recognize a mail
easily ? It would be easier to archive and filter.

Marc


-------------- next part --------------
A non-text attachment was scrubbed...
Name: mjacquin.vcf
Type: text/x-vcard
Size: 416 bytes
Desc: Card for Marc JACQUIN - GEOSYS Inc - Minneapolis 
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991014/e35ec8c1/attachment.vcf>

From TBanister at SpaceImaging.com  Thu Oct 14 09:01:55 1999
From: TBanister at SpaceImaging.com (Banister Trevor)
Date: Thu, 14 Oct 1999 10:01:55 -0600
Subject: [MAPSERVER] Message headers
Message-ID: <AEEF94298C10D311B5A000A024B359E1686D06@imager9.spaceimaging.com>

The common tag is already available. The To: line :-)
Filter on the expression mapserver-users at lists.gis.umn.edu
I'm using this to drop all the mapserver traffic into it's own folder.

> ----------
> From: 	Marc JACQUIN - GEOSYS Inc -
> Minneapolis[SMTP:mjacquin at geosys-inc.com]
> Sent: 	Thursday, October 14, 1999 11:05 AM
> To: 	MapServer Users List
> Subject: 	[MAPSERVER] Message headers
> 
> <<File: mjacquin.vcf>>
> Could we have a common tag inside message headers so we recognize a mail
> easily ? It would be easier to archive and filter.
> 
> Marc
> 
> 
> 


From steve.lime at dnr.state.mn.us  Thu Oct 14 11:10:57 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Thu, 14 Oct 1999 13:10:57 -0500
Subject: Colr look up table
Message-ID: <s805d66d.071@smtp.dnr.state.mn.us>

I assume you're refering to lookup tables based on some attribute value.
Color tables, symbol tables etc. are now handled as classes. Like the old color
table each class has an (optional) expression associated with it. Instead of just
being able to control color, with classes you can control size, color, outlinecolor,
labels and so on... Here's an example with 2 classes :

# what item to I want to classify on
CLASSITEM foo 

# now the class definitions
CLASS
  NAME class1
  EXPRESSION "0"     # note: this is a regular expression
  COLOR 255 0 0
END

CLASS
  NAME class2
  EXPRESSION "1"
  COLOR 0 0 255
END

Assuming item "foo" contains either 0 or 1, 0's would be drawn in red and 1's
in green. Class names are used for legend building, ommit them if you don't
want a particular class to show up in a legend. Features that don't match any
class expression are not drawn. If all features are to be drawn the same, then
define only 1 class with NO expression defined.

Classes are a simple way to create thematic maps based on regular expressions.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

>>> "Paul G. Allen" <pgallen at randomlogic.com> 10/13 4:46 PM >>>
On the ForNet site I was reading some old MapServer documentation (v2.3) about
making a color table. Does the latest version of MapServer support this? I have
a file as outlined in the afore mentioned documentation and I have used the old
printcolors utility to create a GIF image of it. Now I'm wondering if/how I can
use the color file.

Thanks,

PGA




From cshorter at optusnet.com.au  Thu Oct 14 13:33:32 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Fri, 15 Oct 1999 06:33:32 +1000
Subject: [MAPSERVER] Message headers
References: <38060D4E.1A4F8C8F@geosys-inc.com>
Message-ID: <38063E1C.199BFE7D@optusnet.com.au>


Yes, I agree I would find it handy.  If you are using majordomo, then in the
mapserver.config file set the following:
subject_prefix      =   [mapserver]


Marc JACQUIN - GEOSYS Inc - Minneapolis wrote:
> 
> Could we have a common tag inside message headers so we recognize a mail
> easily ? It would be easier to archive and filter.
> 
> Marc


From piero at innova.it  Fri Oct 15 01:45:45 1999
From: piero at innova.it (Pierluigi Vittori)
Date: Fri, 15 Oct 1999 10:45:45 +0200
Subject: Shortest path problem
Message-ID: <99101511024500.00614@buzz.myhouse.net>

Hi all!

I would like to have a new function in mapserv (or as a separate cgi
application) for finding the shortest path between two given points in a city
map.
Does anyone have suggestions - either theoretical or practical - on a
possible approach to this matter?
For example, is Dijkstra's algorithm applicable and suitable to this goal or
are there better ways to accomplish it?

Thanks in advance for your opinions.

Piero


From eserte at cs.tu-berlin.de  Fri Oct 15 14:15:12 1999
From: eserte at cs.tu-berlin.de (Slaven Rezic)
Date: 15 Oct 1999 23:15:12 +0200
Subject: [Fwd: Shortest path problem]
In-Reply-To: Cameron Shorter's message of Fri, 15 Oct 1999 20:48:04 +1000
References: <38070664.A56BEE78@optusnet.com.au>
Message-ID: <87hfjsjpgf.fsf@cabulja.herceg.de>

Hello Cameron, hello Pierluigi,

I use the algorithm A* to get the shortest path between two places. I
don't know what's the format of the mapserver data, but I think it
should be easy to make it usable by my perl module. The module is part
of a larger system. The sources can be found under
	http://pub.cs.tu-berlin.de/src/BBBike

Regards,
	Slaven

-- 
use Tk;$c=tkinit->Canvas(-he,20)->grid;$x=5;map{s/\n//g;map{$c->create('line'=>
map{$a=-43+ord;($x+($a>>3)*2=>5+($a&7)*2)}split//)}split/!/;$x+=12}split/_/=>'K
PI1_+09IPK_K;-OA1_+K!;A__1;Q!7G_1+QK_3CLPI90,_+K!;A_+1!KQ!.N_K+1Q!.F_1+KN.Q__1+
KN._K+1Q!.F_1+KN.Q_+1Q__+1!KQ!.N_1;Q!7G_K3,09Q_+1!K.Q_K+1Q!.F_1+KN.Q_';MainLoop



From piero at innova.it  Sun Oct 17 02:14:24 1999
From: piero at innova.it (Pierluigi Vittori)
Date: Sun, 17 Oct 1999 11:14:24 +0200
Subject: [Fwd: Shortest path problem]
References: <87hfjsjpgf.fsf@cabulja.herceg.de>
Message-ID: <99101711212500.00444@buzz.myhouse.net>

On ven, 15 ott 1999, Slaven Rezic wrote:

> Hello Cameron, hello Pierluigi,
> 
> I use the algorithm A* to get the shortest path between two places. I
> don't know what's the format of the mapserver data, but I think it
> should be easy to make it usable by my perl module. The module is part
> of a larger system. The sources can be found under
> 	http://pub.cs.tu-berlin.de/src/BBBike

Hi Slaven!

I have just taken a look at the address you sent and things look VERY
interesting!
I am going to study it more carefully but for now I just wanted to say thank
you very much for your help.

Regards,
Pierluigi


From pgallen at randomlogic.com  Mon Oct 18 23:13:51 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 18 Oct 1999 23:13:51 -0700
Subject: Map Test Project update
Message-ID: <380C0C1F.9E468A92@randomlogic.com>

I have written a utility specifically designed to concatinate shapefiles using
the Shapelib library called combine. I also had a fellow Linux user write a Perl
script to parse a directory tree, retrieve all files matching a certain pattern,
and run them through the combine program. I've used these two utilities to
combine all county data for two states, and make the states of Illinois and
Pennsylvania on my test web site (http://24.5.14.144) each into their own
shapefile. These two files initially only have county polygon data, but will
later have major roads and water polygon data. Also, since I mucked up the
conversion of the original data from TIGER/Line format to ESRI shapefile format,
I have to do it again to get the actual county names in the database. (ooppss!
:)

It looks as though MapServer, combined with Shapelib, a little Perl, some CGI,
and a yet to be found VRP will give me everything we need for our application.
Now if only I had more HDD storage for all this data! :)

Writing the utility has given me a very good understanding of shapefiles and the
Shapelib library. In fact, I found and fixed a couple of bugs in Shapelib that
bit me more than once. I'll be posting the utilities somewhere and making the
Shapelib author aware of them as well.

Anyone know where I can get an ESRI shapefile of Brazil that includes roads,
landmarks, etc.?

PGA


From bfraser at geoanalytic.com  Tue Oct 19 07:08:54 1999
From: bfraser at geoanalytic.com (Brent Fraser)
Date: Tue, 19 Oct 1999 08:08:54 -0600
Subject: Map Test Project update
References: <380C0C1F.9E468A92@randomlogic.com>
Message-ID: <000e01bf1a3b$7a43ce80$390002c0@servo>

There is some E00 format data at 1:1m at
the GIS Data Depot (probably from the DCW).

http://www.gisdatadepot.com/data/catalog/BR/sublist.html

You'll have to run ESRI's Import71 utility (comes with
ArcView) to convert it to shapefiles...


Brent

----- Original Message -----
From: Paul G. Allen <pgallen at randomlogic.com>
To: MapServer Users List <mapserver-users at lists.gis.umn.edu>
Sent: Tuesday, October 19, 1999 12:13 AM
Subject: Map Test Project update


> I have written a utility specifically designed to concatinate shapefiles
using
> the Shapelib library called combine. I also had a fellow Linux user write
a Perl
> script to parse a directory tree, retrieve all files matching a certain
pattern,
> and run them through the combine program. I've used these two utilities to
> combine all county data for two states, and make the states of Illinois
and
> Pennsylvania on my test web site (http://24.5.14.144) each into their own
> shapefile. These two files initially only have county polygon data, but
will
> later have major roads and water polygon data. Also, since I mucked up the
> conversion of the original data from TIGER/Line format to ESRI shapefile
format,
> I have to do it again to get the actual county names in the database.
(ooppss!
> :)
>
> It looks as though MapServer, combined with Shapelib, a little Perl, some
CGI,
> and a yet to be found VRP will give me everything we need for our
application.
> Now if only I had more HDD storage for all this data! :)
>
> Writing the utility has given me a very good understanding of shapefiles
and the
> Shapelib library. In fact, I found and fixed a couple of bugs in Shapelib
that
> bit me more than once. I'll be posting the utilities somewhere and making
the
> Shapelib author aware of them as well.
>
> Anyone know where I can get an ESRI shapefile of Brazil that includes
roads,
> landmarks, etc.?
>
> PGA
>



From pgallen at randomlogic.com  Tue Oct 19 11:01:56 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Tue, 19 Oct 1999 11:01:56 -0700
Subject: Map Test Project update
References: <380C0C1F.9E468A92@randomlogic.com> <000e01bf1a3b$7a43ce80$390002c0@servo>
Message-ID: <380CB214.EF0A1E80@randomlogic.com>

Brent Fraser wrote:

> There is some E00 format data at 1:1m at
> the GIS Data Depot (probably from the DCW).
>
> http://www.gisdatadepot.com/data/catalog/BR/sublist.html
>
> You'll have to run ESRI's Import71 utility (comes with
> ArcView) to convert it to shapefiles...
>

Now if only I had ArcView...

PGA





From sphillip at mail.esrp.org  Tue Oct 19 12:01:20 1999
From: sphillip at mail.esrp.org (Scott Phillips)
Date: Tue, 19 Oct 1999 12:01:20 PDT
Subject: Map Test Project update (e00 files)
Message-ID: <199910191901.MAA15355@mail.esrp.org>


Re: utilities for e00 files.

I haven't used it myself, but I noticed that there is a utility
called e002gen available from the US Bureau of Land Management.
(ftp://ftp.blm.gov/pub/gis/)  to convert E00 files to Generate
files (ASCII text files that store coordinates)

The index for the directory notes that the utility only works on
line files.  

It looks like it is available as both source code (e002gen.zip) 
and as a DOS executeable (e002gen.exe).

This site also has a copy of shapelib distribution that includes a
program called gen2shp to create a shape file from a generate file.  
(ftp://ftp.blm.gov/pub/gis/shape/)

Note also that the Import71 utility (windows app) can be downloaded 
for free from ESRI without buying ArcView.  However, this utility
does not create shape files, it creates ARC/INFO coverages that are
used with the more expensive ARC/INFO GIS software, but can be read
by ArcView too.  (http://www.esri.com/company/free.html)

If you have GRASS, you could import e00 files and probably export the
data into somthing you could turn into a shape file using shapelib.

Regards,

Scott Phillips
sphillip at esrp.org


> 
> Brent Fraser wrote:
> 
> > There is some E00 format data at 1:1m at
> > the GIS Data Depot (probably from the DCW).
> >
> > http://www.gisdatadepot.com/data/catalog/BR/sublist.html
> >
> > You'll have to run ESRI's Import71 utility (comes with
> > ArcView) to convert it to shapefiles...
> >
> 
> Now if only I had ArcView...
> 
> PGA
> 
> 
> 
> 



From piero at innova.it  Tue Oct 19 11:55:34 1999
From: piero at innova.it (Pierluigi Vittori)
Date: Tue, 19 Oct 1999 20:55:34 +0200
Subject: Shortest path problem
References: <87hfjsjpgf.fsf@cabulja.herceg.de>
Message-ID: <99101920595400.00401@buzz.myhouse.net>

On ven, 15 ott 1999, Slaven Rezic wrote:

> Hello Cameron, hello Pierluigi,
> 
> I use the algorithm A* to get the shortest path between two places. I
> don't know what's the format of the mapserver data, but I think it
> should be easy to make it usable by my perl module. The module is part
> of a larger system. The sources can be found under
> 	http://pub.cs.tu-berlin.de/src/BBBike

Hello Slaven!

can you indicate any papers or other kind of references to turn to for a
theoretical background on such an issue?
I appreciate your help very much. Thanks.

Regards,
Pierluigi


From nhv at cape.com  Tue Oct 19 12:35:10 1999
From: nhv at cape.com (Norman Vine)
Date: Tue, 19 Oct 1999 15:35:10 -0400
Subject: Map Test Project update (e00 files)
In-Reply-To: <199910191901.MAA15355@mail.esrp.org>
Message-ID: <000001bf1a69$480c09e0$3d37ba8c@nhv>

>
>Re: utilities for e00 files.
>

You might find the Grass import e00 'C' code useful
http://www.geog.uni-hannover.de/grass/projects/m.in.e00/welcome.html

Norman Vine





From eserte at cs.tu-berlin.de  Tue Oct 19 16:40:00 1999
From: eserte at cs.tu-berlin.de (Slaven Rezic)
Date: 20 Oct 1999 01:40:00 +0200
Subject: Shortest path problem
In-Reply-To: Pierluigi Vittori's message of Tue, 19 Oct 1999 20:55:34 +0200
References: <87hfjsjpgf.fsf@cabulja.herceg.de>
	<99101920595400.00401@buzz.myhouse.net>
Message-ID: <87bt9u3oof.fsf@cabulja.herceg.de>

Pierluigi Vittori <piero at innova.it> writes:

> 
> On ven, 15 ott 1999, Slaven Rezic wrote:
> 
> > Hello Cameron, hello Pierluigi,
> > 
> > I use the algorithm A* to get the shortest path between two places. I
> > don't know what's the format of the mapserver data, but I think it
> > should be easy to make it usable by my perl module. The module is part
> > of a larger system. The sources can be found under
> > 	http://pub.cs.tu-berlin.de/src/BBBike
> 
> Hello Slaven!
> 
> can you indicate any papers or other kind of references to turn to for a
> theoretical background on such an issue?
> I appreciate your help very much. Thanks.
> 

Here are some:

Kanal,  L.; and Kumar, V., Search in Artificial Intelligence,
Springer-Verlag New York, 1988.

Korf, Richard E., Optimal Path-Finding Algorithms, in Search
in Artificial Intelligence, Springer-Verlag New York, 1988, pages
233-241.

Below is a sample implementation of the algorithm in perl.

Regards,
	Slaven

######################################################################
# -*- perl -*-

#
# $Id: BBBikeDiplom.pm,v 1.3 1999/08/26 19:51:59 eserte Exp $
# Author: Slaven Rezic
#
# Copyright (C) 1999 Slaven Rezic. All rights reserved.
# This package is free software; you can redistribute it and/or
# modify it under the same terms as Perl itself.
#
# Mail: eserte at cs.tu-berlin.de
# WWW:  http://user.cs.tu-berlin.de/~eserte/
#

package BBBikeDiplom;

use Strassen;

package StrassenNetz;

# implementiert die Suche mit dem Algorithmus A*
sub search_A_star {
    my($self, $from, $to) = @_;

    if ($data_format != 1) {
	die "Only implemented for data format 1";
    }

    my $net = $self->{Net};

    # %OPEN entspricht der Liste OPEN aus der Algorithmusbeschreibung
    # Der Key des Hashs ist der Knoten n (die Koordinate im
    # Stra?ennetz), der Value ist ein Array, wovon das erste Element
    # den Vorg?nger enth?lt (oder undef) und das zweite Element die
    # bisherige L?nge g(n).

    # Der Startknoten wird sofort in OPEN eingef?gt (Punkt 1)
    my %OPEN = ($from => [undef, 0]);

    my %CLOSED;

    while (1) {

	# Punkt 2: es gibt keine L?sung
	if (keys %OPEN == 0) {
	    return ();
	}

	# Punkt 3
	my(@sort_list) = sort {
	    # nach Minimum Sortieren
	    $a->[0] <=> $b->[0]
	} map {
	    # f = g + h berechnen
	    my $f = $OPEN{$_}->[1] + Strassen::Util::strecke_s($_, $to);
	    [$f, $_];
	} keys %OPEN;
        # der Knoten (Koordinaten) des Minimums
	my $min_node = $sort_list[0]->[1];
        # ... wird aus OPEN nach CLOSED bewegt
	$CLOSED{$min_node} = $OPEN{$min_node};
	delete $OPEN{$min_node};

	# Punkt 4
	if ($min_node eq $to) {
	    my @path;
	    my $len = 0;
	    while (1) {
		push @path, $min_node;
		my $prev_node = $CLOSED{$min_node}->[0];
		if (defined $prev_node) {
		    $len += Strassen::Util::strecke_s($min_node, $prev_node);
		    $min_node = $prev_node;
		} else {
		    last;
		}
	    }
	    @path = map { [ split(/,/, $_) ] } reverse @path;
	    return (\@path, $len, 0, 0, 0);
	}

	# Punkt 5
	my @successors = keys %{ $net->{$min_node} };
	foreach my $successor (@successors) {
	    my $g = $CLOSED{$min_node}->[1] +  # bisherige L?nge bis n
	      $net->{$min_node}{$successor}; # L?nge bis n'
	    my $f = $g + Strassen::Util::strecke_s($min_node, $to); # h(n')
	    if (!exists $OPEN{$successor} and
		!exists $CLOSED{$successor}) {
		$OPEN{$successor} = [$min_node, $g];
	    } else {
		my $OPEN_OR_CLOSED;
		if (exists $OPEN{$successor}) {
		    $OPEN_OR_CLOSED = $OPEN{$successor};
		} else {
		    $OPEN_OR_CLOSED = $CLOSED{$successor};
		}
		if ($f < $OPEN_OR_CLOSED->[1]) {
		    $OPEN_OR_CLOSED = [$f, $min_node];
		    if (exists $CLOSED{$successor}) {
			$OPEN{$successor} = $CLOSED{$successor};
			delete $CLOSED{$successor}; # hier oder immer? XXX
		    }
		}
	    }
	}
    }
}

1;

__END__
######################################################################

-- 
use Tk;$c=tkinit->Canvas(-he,20)->grid;$x=5;map{s/\n//g;map{$c->create('line'=>
map{$a=-43+ord;($x+($a>>3)*2=>5+($a&7)*2)}split//)}split/!/;$x+=12}split/_/=>'K
PI1_+09IPK_K;-OA1_+K!;A__1;Q!7G_1+QK_3CLPI90,_+K!;A_+1!KQ!.N_K+1Q!.F_1+KN.Q__1+
KN._K+1Q!.F_1+KN.Q_+1Q__+1!KQ!.N_1;Q!7G_K3,09Q_+1!K.Q_K+1Q!.F_1+KN.Q_';MainLoop



From cshorte2 at csc.com  Tue Oct 19 18:47:18 1999
From: cshorte2 at csc.com (cshorte2 at csc.com)
Date: Wed, 20 Oct 1999 11:47:18 +1000
Subject: Shortest path problem (translating from German)
Message-ID: <85256810.000ACA12.00@csc.com>



If you are having problems understanding the German comments you can translate
it at:
http://babelfish.altavista.digital.com/




From pgallen at randomlogic.com  Fri Oct 22 18:01:51 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 22 Oct 1999 18:01:51 -0700
Subject: LArge area mapping problem.
Message-ID: <381108FF.303D3020@randomlogic.com>

I'm having a problem navigating larger maps. I'm not sure if it's my data or
what. I have a map of Pennsylvania that was made by combining all the TIGER
county shapefiles into a single shapefile. Displaying the initial map looks OK,
with one exception: the county names are displayed off the top of the map (out
of the visible map area). An additional problem comes into effect when
panning/zooming. Clicking anywhere in the map for a pan or zoom does not zoom
into the location clicked, but off to the lower right by some amount. Also, if I
set the Zoom to 1, and click the "Refresh/Query" button, instead of the map
simply refreshing what is already there, it pans in some direction (I think down
and right) by some amount.

I'm using MapServer 3.20, and to see what it's doing for yourself, visit
http://24.5.14.144. I've checked my map extents and they look to be OK (they are
in DD, same as the original data).

PGA


From pgallen at randomlogic.com  Fri Oct 22 18:06:46 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 22 Oct 1999 18:06:46 -0700
Subject: LArge area mapping problem.
References: <381108FF.303D3020@randomlogic.com>
Message-ID: <38110A26.921C2A30@randomlogic.com>

In the last paragraph, I said "I'm using MapServer 3.20...", I meant MapServer
3.2.20. I haven't compiled 3.2.21 yet, though I did D/L it.

PGA

"Paul G. Allen" wrote:
> 
> I'm having a problem navigating larger maps. I'm not sure if it's my data or
> what. I have a map of Pennsylvania that was made by combining all the TIGER
> county shapefiles into a single shapefile. Displaying the initial map looks OK,
> with one exception: the county names are displayed off the top of the map (out
> of the visible map area). An additional problem comes into effect when
> panning/zooming. Clicking anywhere in the map for a pan or zoom does not zoom
> into the location clicked, but off to the lower right by some amount. Also, if I
> set the Zoom to 1, and click the "Refresh/Query" button, instead of the map
> simply refreshing what is already there, it pans in some direction (I think down
> and right) by some amount.
> 
> I'm using MapServer 3.20, and to see what it's doing for yourself, visit
> http://24.5.14.144. I've checked my map extents and they look to be OK (they are
> in DD, same as the original data).
> 
> PGA


From cshorter at optusnet.com.au  Sun Oct 24 01:49:08 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Sun, 24 Oct 1999 18:49:08 +1000
Subject: Speeding up mapping applications by changing data structures
References: <s7e2528f.000@smtp.dnr.state.mn.us> <37E2AF6E.C44205EC@home.com> <37F723AB.93350DDA@optusnet.com.au> <37F78890.291DC130@home.com>
Message-ID: <3812C804.1FF2CBCC@optusnet.com.au>

I recently heard a speach from one of the silicon graphics guys, who
explained, among other things, where technology is heading, and what that
means for writing programs.

CPU instruction speeds are increasing faster than RAM access speeds.  Hence,
with old machines, an access to RAM took one CPU cycle, now RAM accesses take
lots of CPU cycles.
Also, it is faster to access a block of RAM than it is to access the same
ammount of data from a number of different locations, partly because CPUs now
make use of memory look ahead.

What this means, is that programs should use pointers sparingly, and I belive
this explains Steve's comment that the Shapelib 1.2 algorithms are
significantly slower than Shapelib 1.1.

The shape object in shapefil.h is copied below:

typedef struct
{
    int		nSHPType;

    int		nShapeId; /* -1 is unknown/unassigned */

    int		nParts;
    int		*panPartStart;
    int		*panPartType;
    
    int		nVertices;
    double	*padfX;
    double	*padfY;
    double	*padfZ;
    double	*padfM;

    double	dfXMin;
    double	dfYMin;
    double	dfZMin;
    double	dfMMin;

    double	dfXMax;
    double	dfYMax;
    double	dfZMax;
    double	dfMMax;
} SHPObject;

Note that the X,Y,Z,M coordinates are in seperate NULL terminated lists.  When
a vertex is accessed, 4 seperate pointers need to be resolved.  The ESRI
shapefile record stucture stores the data much more efficently as
(X,Y,Z,M),(X,Y,Z,M),...

Steve,
Your structures for the next release are better, however I belive they can be
improved also.  As I understand it, an object is of the form:
  (line1, line2, line3, ...)
which expands out to:
  (nPoints,(x,y),(x,y),..) , (nPoints,(x,y),(x,y),..), ...

I suspect you could improve efficiency slightly if the "nPoints" is moved out
into another array similar to the ESRI structure, although improvements would
probably be marginal.

Steve and Frank,
To help the structure of supporting packages, it would be good if coordinates
are writting as an array - ( c[0],c[1],c[2],c3 ) instead of (x,y,z,m).

It would then be easier to write generic procedures which can handle 2, 3, or
4 dimentional coordinates.

Lastly,
I realise that data structures effects everything that depends upon it, and
hence a lot of work, so I'm not expecting you to change everything.


From ddnebert at usgs.gov  Mon Oct 25 11:55:46 1999
From: ddnebert at usgs.gov (Doug Nebert)
Date: Mon, 25 Oct 1999 14:55:46 -0400
Subject: Using Java with mapserver
Message-ID: <3814A7B2.5178867@usgs.gov>

I have a need to build a lightweight java applet that will permit 
pan and zoom into a certain extent, then allow the user to drag a 
rectangle (yes, java can do this!) to set an area of interest. The 
result of all of this is a bit anticlimactic: I only need the 
transformed coordinates of the rectangle fed to a containing HTML 
as shared variables. I have a program that uses LiveConnect to pass 
the java variables to the hosting form. But I was wondering if 
someone else has some java stubs for making draw, pan, and zoom 
requests of a mode=map service...

-- 
Douglas D. Nebert
Clearinghouse Coordinator 
FGDC Secretariat	Phone: +1 703 648 4151	Fax: +1 703 648-5755	
Pager Messaging: 	http://fgdclearhs.er.usgs.gov/dougmsg.html


From steve.lime at dnr.state.mn.us  Mon Oct 25 12:26:23 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Mon, 25 Oct 1999 14:26:23 -0500
Subject: Using Java with mapserver
Message-ID: <s81468af.090@smtp.dnr.state.mn.us>

I've already done this. I have a little LiveConnect applet that 
is used to do box zooms and/or selects. See the data deli
(http://deli.dnr.state.mn.us/) for an example. On the deli the
user pans/zooms using box or point/zoomsize controls and
can go into a query mode that uses the same mechanism to
create a query. If this looks like what you're interested in then
I'll post the java code and instructions for use. Nice thing is that
the applet is a whopping 6k in size, nice and small.

Note: the deli uses public methods within the applet to toggle
box creation on/off. This makes the interface a little cleaner.
So, for example, when zooming out you don't want to be able
to drag a box.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

>>> Doug Nebert <ddnebert at usgs.gov> 10/25 1:55 PM >>>
I have a need to build a lightweight java applet that will permit 
pan and zoom into a certain extent, then allow the user to drag a 
rectangle (yes, java can do this!) to set an area of interest. The 
result of all of this is a bit anticlimactic: I only need the 
transformed coordinates of the rectangle fed to a containing HTML 
as shared variables. I have a program that uses LiveConnect to pass 
the java variables to the hosting form. But I was wondering if 
someone else has some java stubs for making draw, pan, and zoom 
requests of a mode=map service...

-- 
Douglas D. Nebert
Clearinghouse Coordinator 
FGDC Secretariat	Phone: +1 703 648 4151	Fax: +1 703 648-5755	
Pager Messaging: 	http://fgdclearhs.er.usgs.gov/dougmsg.html



From pgallen at randomlogic.com  Mon Oct 25 20:07:12 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 25 Oct 1999 20:07:12 -0700
Subject: I don't get it, Penn. is still messed up.
Message-ID: <38151AE0.24C50C60@randomlogic.com>

After fixing the inconsistency between the image size between the template file
and the map file, the pan/zoom now seems to work fine. The county names are
still FUBAR though. I added some nice large point symbols to indicate where
MapServer was plotting the county name points themselve, and they are all
plotting correctly. The labels however, are FUBAR - they are all WAY above the
map, and begin to float down as you zoom in. What's up with that?

PGA


From pgallen at randomlogic.com  Mon Oct 25 20:37:00 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Mon, 25 Oct 1999 20:37:00 -0700
Subject: Yet another small map utility
Message-ID: <381521DC.B7FB8006@randomlogic.com>

I wrote a small map utility using the shapelib API for creating shapefiles (yes,
I know, there are several already, but none were exactly what I wanted). This
particular utility takes several command line arguments and makes a complete
POINT shapefile from them. If you look at my Philadelphia map, you'll see a
couple red SIZE 10 points that were created by this little utility. The command
line format is:

[root at igen /root]# $SHAPELIB/unitoshape
unitoshape [path]<output_file> shapetype LAT LONG ID equip_type msg_time
[string_array] 
    Where string_array is:
        <DBBASE_field_name> <field_type(N|C)> <width> <precision> <field_data>

    string_array may be repeated for as many DBASE fields as necessary,
    up to the max argv[] limit of the OS.

(NOTE: $SHAPELIB is an env. variable that makes things easier for me, it's not
required for the utility. It just points to my shapelib development dir.)

[path]<output_file> = the path and file name of the shapefile to create or
append to (no extension, the utility adds them)
shapetype = The shape type to create. Only supports POINTs for now (that's all I
need right now)
LAT LONG = Dotted decimal latitude and longitude of the shape position
ID = A numerical ID of the shape of type INTEGER
equip_type = A string describing what this shape is (e.g., Warehouse)
msg_time = The time that this shapes data was last updated, a string.

The first seven command line options are required, but the DBASE fields are not.
The reason I wrote this is to take data from a number of different sources, in
any number of different formats, and plot them on a map, as well as create DBASE
files that could be queried by MapServer. A couple simple Perl scripts can (and
will) be written to extract data from an SQL database, an email message, or
whatever, and pipe the data to the unitoshape utility for quick and painless
plotting. Basically, I have a universal interface for data extraction and
insertion with the ability to support additional data formats quickly - maybe
today I have the data in a SQL database, and tomorrow I have it in simple text
files.

If anyone would like this utility, and a couple Perl scripts to go with it, I'll
be posting them soon. I want to add a file locking mechanism first so that the
same shapefile won't be updated by two unitoshape processes at the same time. I
also plan to make the script create a backup of the original shapefile during
updates to prevent corruption and web page access delays when a web browser
tries to access a shapefile that's being updated.

PGA


From steve.lime at dnr.state.mn.us  Tue Oct 26 08:24:08 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 26 Oct 1999 10:24:08 -0500
Subject: Speeding up mapping applications by changing data
	structures
Message-ID: <s8158165.043@smtp.dnr.state.mn.us>

Your points are well taken. I chose the more structured approach for a
couple of reasons:

  - readablity. I actually made the jump from the ESRI xyxyxyxy way of
    doing things when implementing new clipping functionality in version
    3.2. Working with the xyxyxyxy list isn't hard, but dealing with parts
    arrays was nasty. The code now is a good bit more readable and is
    much easier to re-visit after a few months.

  - compatibility. I've always envisioned tapping in to OGC or OGDI
    resources as a means to extend the MapServer. It's pretty clear that
    higher level abstractions of a feature would be necessary to make
    that integration simpler.

I've contemplated the switch you outline below a couple of times, but haven't
done it. What really changes is the definition of a point- from a structure to an
2 element array. It could be done, a couple hundred changes, but all straight
forward. Probably could do it with 2 sed commands. I'd be interested in the
performance benefits.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

>>> Cameron Shorter <cshorter at optusnet.com.au> 10/24 3:49 AM >>>
I recently heard a speach from one of the silicon graphics guys, who
explained, among other things, where technology is heading, and what that
means for writing programs.

CPU instruction speeds are increasing faster than RAM access speeds.  Hence,
with old machines, an access to RAM took one CPU cycle, now RAM accesses take
lots of CPU cycles.
Also, it is faster to access a block of RAM than it is to access the same
ammount of data from a number of different locations, partly because CPUs now
make use of memory look ahead.

What this means, is that programs should use pointers sparingly, and I belive
this explains Steve's comment that the Shapelib 1.2 algorithms are
significantly slower than Shapelib 1.1.

The shape object in shapefil.h is copied below:

typedef struct
{
    int		nSHPType;

    int		nShapeId; /* -1 is unknown/unassigned */

    int		nParts;
    int		*panPartStart;
    int		*panPartType;
    
    int		nVertices;
    double	*padfX;
    double	*padfY;
    double	*padfZ;
    double	*padfM;

    double	dfXMin;
    double	dfYMin;
    double	dfZMin;
    double	dfMMin;

    double	dfXMax;
    double	dfYMax;
    double	dfZMax;
    double	dfMMax;
} SHPObject;

Note that the X,Y,Z,M coordinates are in seperate NULL terminated lists.  When
a vertex is accessed, 4 seperate pointers need to be resolved.  The ESRI
shapefile record stucture stores the data much more efficently as
(X,Y,Z,M),(X,Y,Z,M),...

Steve,
Your structures for the next release are better, however I belive they can be
improved also.  As I understand it, an object is of the form:
  (line1, line2, line3, ...)
which expands out to:
  (nPoints,(x,y),(x,y),..) , (nPoints,(x,y),(x,y),..), ...

I suspect you could improve efficiency slightly if the "nPoints" is moved out
into another array similar to the ESRI structure, although improvements would
probably be marginal.

Steve and Frank,
To help the structure of supporting packages, it would be good if coordinates
are writting as an array - ( c[0],c[1],c[2],c3 ) instead of (x,y,z,m).

It would then be easier to write generic procedures which can handle 2, 3, or
4 dimentional coordinates.

Lastly,
I realise that data structures effects everything that depends upon it, and
hence a lot of work, so I'm not expecting you to change everything.



From mdh20 at axe.humboldt.edu  Tue Oct 26 08:47:53 1999
From: mdh20 at axe.humboldt.edu (Michael Hass)
Date: Tue, 26 Oct 1999 08:47:53 -0700 (PDT)
Subject: Using Java with mapserver
In-Reply-To: <s81468af.090@smtp.dnr.state.mn.us>
Message-ID: <Pine.OSF.4.10.9910260844050.1887-100000@axe.humboldt.edu>


I'd be interested to see this posted so that we can try it out on our
project. It brings up another java question, when trying to run Example 5
with the java applet, I get an error that it can't load the applet
mapimage class mapimage. Is this a path problem or do I need the java
code?
Thanks,
Mike Hass
Research Assistant
Klamath River Restoration GIS
Humboldt State University
Arcata, CA

On Mon, 25 Oct 1999, Stephen Lime wrote:

> I've already done this. I have a little LiveConnect applet that 
> is used to do box zooms and/or selects. See the data deli
> (http://deli.dnr.state.mn.us/) for an example. On the deli the
> user pans/zooms using box or point/zoomsize controls and
> can go into a query mode that uses the same mechanism to
> create a query. If this looks like what you're interested in then
> I'll post the java code and instructions for use. Nice thing is that
> the applet is a whopping 6k in size, nice and small.
> 
> Note: the deli uses public methods within the applet to toggle
> box creation on/off. This makes the interface a little cleaner.
> So, for example, when zooming out you don't want to be able
> to drag a box.
> 
> Steve
> 
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
> 
> (651) 297-2937
> steve.lime at dnr.state.mn.us
> 
> >>> Doug Nebert <ddnebert at usgs.gov> 10/25 1:55 PM >>>
> I have a need to build a lightweight java applet that will permit 
> pan and zoom into a certain extent, then allow the user to drag a 
> rectangle (yes, java can do this!) to set an area of interest. The 
> result of all of this is a bit anticlimactic: I only need the 
> transformed coordinates of the rectangle fed to a containing HTML 
> as shared variables. I have a program that uses LiveConnect to pass 
> the java variables to the hosting form. But I was wondering if 
> someone else has some java stubs for making draw, pan, and zoom 
> requests of a mode=map service...
> 
> -- 
> Douglas D. Nebert
> Clearinghouse Coordinator 
> FGDC Secretariat	Phone: +1 703 648 4151	Fax: +1 703 648-5755	
> Pager Messaging: 	http://fgdclearhs.er.usgs.gov/dougmsg.html
> 
> 



From steve.lime at dnr.state.mn.us  Wed Oct 27 11:14:54 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Wed, 27 Oct 1999 13:14:54 -0500
Subject: Java code...
Message-ID: <s816faf2.037@smtp.dnr.state.mn.us>

Ok. Here's my working version. It's Java 1.1 code and compiles fine
under 1.2 except for a few "depricated methods" messages. If anyone
is interested in porting to 1.2 and swing let me know.

To install: Copy the mapimage.class file someplace on you system that
is web accessible. If you need to rebuild from source you'll need to make
sure that the Netscape java classes (either a .zip or a .jar) are in your
CLASSPATH environment variable.

To use:

  1. Make sure your normal MapServer form has a name attribute (eg. name="mapserv")
  2. Add the following javascript code to the HTML header section:

    <script language="javascript">
      function SetImageXY(i) { document.mapserv.imgxy.value = i; }
      function SetImageBox(i) { document.mapserv.imgbox.value = i; }
    </script>

     These functions are called by the applet, so the names can't be changed. However, you
     could change the function code itself to, for example:
    
     - submit the form immediately
     - update the status line of the browser

  3.  Add a couple of hidden variables to the form:

    <input type="hidden" name="imgbox" value="-1 -1 -1 -1">
    <input type="hidden" name="imgxy" value="149.5 199.5">
  
    imgbox allows the MapServer to process a pair of mouse clicks (ie. a box). It is ignored if the box is 
    simply a point, hence the initialization value. imgxy is a way to pass a single mouse click to the
    MapServer. This mimics the standard forms based method. The initial value should be something
    meaningful like the center of the image. Remember that the coordinates on the image start at 0
    and end at nrows-1 or ncols-1. To compute center x use (ncols-1)/2 and y use (nrows-1)/2 and
    keep the decimal value. This allows the user to zoom using a button rather than the image.

  4. Add a call to the applet (this replaces the normal form-based call to the image <input type="image"...>).
      Here's an example from a template file (note the [img] substitution):

    <applet codebase="/java/mapimage" code="mapimage.class" width="300" height ="400" name="mapimage" MAYSCRIPT>
        <param name="image" value="http://www.dnr.state.mn.us[img]">
        <param name="jitter" value="1">
        <param name="color" value="255 0 0">
      </applet>   

     The parameters are:

       image - fully qualified URL to the image. This can be a CGI call.
       jitter - allowable slop in mouse clicking. Basically sets a minimum box size.
       color - RGB triplet that sets the color of the box or cursor.

That covers the basic use. There are 3 public methods that can be called from javascript that change
the behavior of the applet.

boxOn and boxOff toggle the applets ability to display a dragged box. Useful when adding features to
an interface. For example, it doesn't make sense to drag a box when zooming out. No arguments.

newMap(string image_url) allows you to swap images without changing the rest of the page. The function
uses the java MediaTracker interface to delay image loading until the new image is completely loaded. This
is useful with advanced interface setup. It's possible to manage extents, layers etc completely on the client
using javascript requesting a new image only when necessary via the applet. It's actually pretty cool once
it's set up that way. The string supplied to newMap must be a fully qualified URL for an image (can be a CGI
call).

Examples of this applet can be found at:

  - http://deli.dnr.state.mn.us/
  - http://www.dnr.state.mn.us/mis/gis/gislis99/example1.html (click "next step" until you hit example 5)
  - http://www.ra.dnr.state.mn.us/

Possible enhancements:

  - displaying coordinates on the image as the mouse moves
  - bubble attributes (eg. move mouse over feature label and display attributes or whatever)

Hope this helps, have fun!

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

-------------- next part --------------
A non-text attachment was scrubbed...
Name: mapimage.java
Type: application/octet-stream
Size: 4744 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991027/f316d528/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mapimage.class
Type: application/octet-stream
Size: 5320 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991027/f316d528/attachment-0001.obj>

From steve.lime at dnr.state.mn.us  Thu Oct 28 07:38:55 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Thu, 28 Oct 1999 09:38:55 -0500
Subject: updated java...
Message-ID: <s81819d4.014@smtp.dnr.state.mn.us>

Hi: Here's a new version of the mapimage applet. I just changed the behavior
when the image is loading a bit. Nothing major but it's a bit faster. Note this
is still Java 1.0 API code. The mouse events need to be messed with to make
it 1.1+ compliant. However, the old code works with older browsers...

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us

-------------- next part --------------
A non-text attachment was scrubbed...
Name: mapimage.java
Type: application/octet-stream
Size: 4731 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991028/7714a1db/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mapimage.class
Type: application/octet-stream
Size: 5292 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19991028/7714a1db/attachment-0001.obj>

From ddnebert at usgs.gov  Thu Oct 28 13:39:47 1999
From: ddnebert at usgs.gov (Doug Nebert)
Date: Thu, 28 Oct 1999 16:39:47 -0400
Subject: Java Script questions
Message-ID: <3818B493.A0CA7B65@usgs.gov>

I have been attempting to apply the mapimage.java implementation 
and have been having a number of problems: mostly due to lack of 
proficiency with JavaScript and the exchange of information between 
HTML and the applet. 

Current implementation: http://clearinghouse3.fgdc.gov/global.html

The short explanation is that I cannot get the map to zoom. If I 
use a setting as suggested: 

<param name="image" value="http://clearinghouse3.fgdc.gov[img]">

First of all I don't see the map, just a big blue image. Which implies 
that I am actually seeing the map, but am stuck out in the ocean 
at some pixel coordinates instead of at some "geographic" coordinates.
I notice that if I use the rectangle box to zoom it is somehow setting 
&imgbox=102+45+270+143 in the GET call I see, but that doesn't 
mean much in the registered image coordinates.  I need to be able 
to translate the pixel coordinates into the "lat-long" coordinates 
1) to pass to the mapserver to help formulate a &mapext and 2) to 
report the rectangle's coordinates out to the HTML form using it 
for other purposes. I am not sure how to approach this and whether/how 
to use Java or JavaScript to enable it. As a side note, I do want to 
be able to use this map selection in the context of a bigger "form" 
to which I can post the coordinates pulled from the map. This may 
suggest that I need to use Java exclusively as a component in a 
bigger form, right?

This will take several steps:

1. Store the current latitude and longitude values for the displayed 
map. On startup these are: latmin = -90, latmax = 90, longmin = -180, 
and longmax = 180 though these could also be set in the calling 
HTML file for different startup regions. Where/how do I do this?

2. Collect a rectangle within the java applet. This appears to set
imgxy and imgbox values as the pixel coordinates measureed from the 
upper left corner of the canvas. This seems to be done already.

3. I think that one would need to calculate the appropriate
map dimensions to grab based on the doodled rectangle. One would 
determine the coordinate bounds of a rectangle that more than 
encloses the selected rectangular region. I will always request a 
map that is twice as wide as it is tall (360 W-E x 180 N-S degrees 
sets my design aspect ratio). This seems to be done instead inside 
the applet:

SetImageBox(\"" + Math.min(x1,x2) + " " + Math.min(y1,y2) + " " 
+ Math.max(x1,x2) + " " + Math.max(y1,y2) + "\"); ");

Doesn't this request pixel coordinates rather than geographic 
coordinates? I need some external control to derive not the 
imgbox but two mapext values and pack them into the GET command:
I need the requested map extent, and I need the selected rectangle 
bounds drawn on it. The transformation math is as follows:

xCoord = longmin + (( x / wide ) * ( longmax - longmin ));
yCoord = latmin + ((( tall - y ) / tall ) * ( latmax - latmin ));

where wide = 500 and tall = 250 as integer pixel values. 

How do I differentiate between the red rectangle coordinates and 
the mapext requesting coordinates? I need to set and manage both.

4. Request a map with the correct latitude and longitude values
expressed in the &mapext settings, and not use "zoom" at all 
via the GET method (or java equivalent).

5. Have the selection rectangle shown in red on the resulting map.

Tutorial advice on how to achieve this would be appreciated, as 
I am sure many of you have dealt with these kind of issues before. 
By your suggestions, I can better understand how this all works.
As I describe it here, I come to the conclusion that I probably 
need a full java applet to do all the interaction I need... Any
suggestions or examples?

-- 
Douglas D. Nebert
Clearinghouse Coordinator 
FGDC Secretariat	Phone: +1 703 648 4151	Fax: +1 703 648-5755	
Pager Messaging: 	http://fgdclearhs.er.usgs.gov/dougmsg.html


From pgallen at randomlogic.com  Thu Oct 28 20:01:57 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Thu, 28 Oct 1999 20:01:57 -0700
Subject: HTML Help
Message-ID: <38190E25.69DA83A5@randomlogic.com>

Thanks to the KPLUG list for the help on password protecting my map web pages.
My San Diego map now requires a password to access. (I needed to do this because
there may be sensitive data on it).

The other pages are still available to the general public though.

PGA


From jBiza at muhb.cz  Thu Oct 28 23:42:17 1999
From: jBiza at muhb.cz (=?windows-1250?Q?B=ED=9Ea_Josef_ing=2E?=)
Date: Fri, 29 Oct 1999 08:42:17 +0200
Subject: Java code...
Message-ID: <BF0859F4090DD311A39600902772C34D02E732@s001.intra.muhb.cz>


Thank you very much Steve.

Josef Biza

Town Hall Havlickuv Brod
Havlickovo namesti 57
580 01 Havlickuv Brod.
Czech Republic.


From cshorter at optusnet.com.au  Fri Oct 29 03:19:20 1999
From: cshorter at optusnet.com.au (Cameron Shorter)
Date: Fri, 29 Oct 1999 20:19:20 +1000
Subject: mapserver modify shape proposal
References: <s8187d34.012@BWSR-STPAUL-NW411>
Message-ID: <381974A8.703D69B6@optusnet.com.au>


Tim Ogg wrote:
> 
> Cameron -
> 
>   I have seen your communication with the mapserver discussion group and Steve Lime about
> adding shape file modification capability to mapeserver. I am not a programmer. I am merely a GIS person
> who has an application that would require shape file editing via a browser using something like mapserver.
> I don't want to pay for something like ESRI-IMS if I don't have to.
> 
>   From the discussion archive it isn't clear if you are actually going to attempt to implement editing
> capabilities or not. Where do you stand on the issue? If you plan to go ahead, do you have an
> intended timeline at all? I know this is most likely a situation where you work on it when you can
> and it is impossible to know the timeline but I thought it couldn't hurt to ask.


Tim,
Yes, I am working on a web based modify-map program.  It will be a seperate
program to mapserver, but should dovetail with it well.  Much as I'd like to
work on this full time, I have to earn my keep as well, so timelimes are
pushed out a bit, but this is a very rough idea of what I'm expecting to have
completed and when:

End of November : Finish cgi parser (I'm half way through this now)
January (ish)   : Allow a user to add a shape to a ERSI object
<release point>
March 2000      : Write some front end application (in java or javascript)
that will allow an operator to add a map.
<release point>
June 2000       : Incorporate Modify/Delete capabilities
September 2000  : Clean up error handling and so forth

This is assuming that no one volunteers to help out.


From bowenj at gaiaenv.com  Fri Oct 29 07:40:59 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Fri, 29 Oct 1999 08:40:59 -0600
Subject: Query template file question
Message-ID: <3819B1FB.CF27D26D@gaiaenv.com>

Hi all,

Probably a question for Steve, but perhaps someone else can help.

We are using PHP3 to access database information via our site.  We are
also using PHP3 to link MapServer with PostgreSQL.  The links are
rudimentary at this point, but show potential.  At this point, all we
have in place is:  Query function passes a unique id to PHP3 script that
then performs a query on PostgreSQL and dynamically creates html
output.  

My question lies in the 'query template' identified in the MapServer map
file.  The docs indicate that output can be passed to an URL or template
file.  I cannot find docs or eg. relating to passing the query info. to
an URL.  I would like to do this because MapServer does not recognise a
.PHP3 extension as a valid html template file.  I think I found where I
would need to change this in the source, but perhaps someone could shed
some light on this (C is not my forte).  I do not want to dick with my
apache settings...I believe source change would be fastest and least
intrusive to our web server.  Currently, I have a temporary html
template that grabs MapServer variables, and passes them via the URL
string, using an http-equiv redirect. (yuck)

I would have included the URL for the site, but the information is
confidential (client request...sorry).  Thanks in advance.


Cheers,
Jim
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA


From bfraser at geoanalytic.com  Fri Oct 29 08:12:44 1999
From: bfraser at geoanalytic.com (Brent Fraser)
Date: Fri, 29 Oct 1999 09:12:44 -0600
Subject: Query template file question
References: <3819B1FB.CF27D26D@gaiaenv.com>
Message-ID: <000901bf2220$0d55d510$390002c0@servo>

Jim,

  I'm not sure what version of MapServer you're running, but if
it's recent, have a look at the mapserv.ini file.  It has an entry
for setting the template extension:

# File extention of template html files. Should be a regular expression.
This
# usually doesn't need to be changed.
TEMPLATE_EXTENSION=\.html$

I'm not sure if this will solve your problem though.  If it doesn't,
perhaps you could explain the flow of control between mapserver
and PHP3 in more detail.

Brent Fraser


----- Original Message -----
From: Jim T. Bowen (GAIA Consultants Inc.) <bowenj at gaiaenv.com>
To: MapServer <mapserver-users at lists.gis.umn.edu>
Sent: Friday, October 29, 1999 8:40 AM
Subject: Query template file question


> Hi all,
>
> Probably a question for Steve, but perhaps someone else can help.
>
> We are using PHP3 to access database information via our site.  We are
> also using PHP3 to link MapServer with PostgreSQL.  The links are
> rudimentary at this point, but show potential.  At this point, all we
> have in place is:  Query function passes a unique id to PHP3 script that
> then performs a query on PostgreSQL and dynamically creates html
> output.
>
> My question lies in the 'query template' identified in the MapServer map
> file.  The docs indicate that output can be passed to an URL or template
> file.  I cannot find docs or eg. relating to passing the query info. to
> an URL.  I would like to do this because MapServer does not recognise a
> .PHP3 extension as a valid html template file.  I think I found where I
> would need to change this in the source, but perhaps someone could shed
> some light on this (C is not my forte).  I do not want to dick with my
> apache settings...I believe source change would be fastest and least
> intrusive to our web server.  Currently, I have a temporary html
> template that grabs MapServer variables, and passes them via the URL
> string, using an http-equiv redirect. (yuck)
>
> I would have included the URL for the site, but the information is
> confidential (client request...sorry).  Thanks in advance.
>
>
> Cheers,
> Jim
>  ___________________________________    GGG G   AAA      III    AAA
> |  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A
> |                                   | G     G    A A      I      A A
> |Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A
> |Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A
> |Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A
> |_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA
>            ICQ: 15567315              G     G  A     A    I    A     A
>       http://www.gaiaenv.com           G    G  A     A    I    A     A
>                                         GGGG  AAA   AAA  III  AAA   AAA
>



From kenboss at dilbert.dnr.state.mn.us  Fri Oct 29 08:31:04 1999
From: kenboss at dilbert.dnr.state.mn.us (kenboss at dilbert.dnr.state.mn.us)
Date: Fri, 29 Oct 1999 10:31:04 -0500 (CDT)
Subject: Query template file question
Message-ID: <199910291531.KAA07517@philbert.dnr.state.mn.us>

Jim--

If I'm understanding your question correctly, all you need to do is to put the URL in quotes after the 
"template" keyword in the query block in your map file.  So, instead of something like this:

LAYER
  ...
  ...
  QUERY
    TEMPLATE "http://www.dnr.state.mn.us/someapp/somefile.html"
  END
END


You have something like this:

LAYER
  ...
  ...
  QUERY
    TEMPLATE "http://www.dnr.state.mn.us/perl/lk_search.pl?var1=[SOMEFIELD]&var2=[ANOTHER_FIELD]"
  END
END


Where [SOMEFIELD] and [ANOTHER_FIELD] are attributes of the shape that the user clicked on (eg., the 
unique id you want to pass to your PHP3 script).

This can have all sorts of cool applications, for example, we use:

LAYER
  NAME counties
  TYPE polyline
  STATUS default
  DATA counties
  CLASS
    NAME "County Boundaries"
    SYMBOL 1
    SIZE 1
    COLOR 0 0 0
  END
  QUERY
    TEMPLATE "http://www.ra.dnr.state.mn.us/cgi-bin/mapserv3.2?map=photos/etc/photos.map&mapext=[shpext]"
  END
END # Layer counties


to allow users to zoom in to the extent of a county from an initial statewide interface (see 
http://www.ra.dnr.state.mn.us/photos/photos_init.html).  Note that in this case [shpext] is not an 
attribute of the shape per se (that is, it's not stored in the .dbf file), but is a value calculated by 
the mapserver software (documented at http://mapserver.gis.umn.edu/templates.html).

Hope this helps,

--Ken


============================================================================
Ken Boss                                
Digital Image Analysis / Web Stuff      Forestry Resource Assessment
kenboss at dilbert.dnr.state.mn.us         Minnesota Dept. of Natural Resources
Voice: 218 327 4449 ext. 237            413 SE 13th Street                    
Fax:   218 327 4517                     Grand Rapids, MN  55744         USA
============================================================================




From bowenj at gaiaenv.com  Fri Oct 29 08:43:10 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Fri, 29 Oct 1999 09:43:10 -0600
Subject: Query template file question
References: <3819B1FB.CF27D26D@gaiaenv.com> <000901bf2220$0d55d510$390002c0@servo>
Message-ID: <3819C08E.276E74B7@gaiaenv.com>

Cheers Brent.  We are running 3.219b...never thought of checking the ini
(still have to get use to that feature!).  I will give it a try later
today.

Jim

Brent Fraser wrote:
> 
> Jim,
> 
>   I'm not sure what version of MapServer you're running, but if
> it's recent, have a look at the mapserv.ini file.  It has an entry
> for setting the template extension:
> 
> # File extention of template html files. Should be a regular expression.
> This
> # usually doesn't need to be changed.
> TEMPLATE_EXTENSION=\.html$
> 
> I'm not sure if this will solve your problem though.  If it doesn't,
> perhaps you could explain the flow of control between mapserver
> and PHP3 in more detail.
> 
> Brent Fraser
> 
> ----- Original Message -----
> From: Jim T. Bowen (GAIA Consultants Inc.) <bowenj at gaiaenv.com>
> To: MapServer <mapserver-users at lists.gis.umn.edu>
> Sent: Friday, October 29, 1999 8:40 AM
> Subject: Query template file question
> 
> > Hi all,
> >
> > Probably a question for Steve, but perhaps someone else can help.
> >
> > We are using PHP3 to access database information via our site.  We are
> > also using PHP3 to link MapServer with PostgreSQL.  The links are
> > rudimentary at this point, but show potential.  At this point, all we
> > have in place is:  Query function passes a unique id to PHP3 script that
> > then performs a query on PostgreSQL and dynamically creates html
> > output.
> >
> > My question lies in the 'query template' identified in the MapServer map
> > file.  The docs indicate that output can be passed to an URL or template
> > file.  I cannot find docs or eg. relating to passing the query info. to
> > an URL.  I would like to do this because MapServer does not recognise a
> > .PHP3 extension as a valid html template file.  I think I found where I
> > would need to change this in the source, but perhaps someone could shed
> > some light on this (C is not my forte).  I do not want to dick with my
> > apache settings...I believe source change would be fastest and least
> > intrusive to our web server.  Currently, I have a temporary html
> > template that grabs MapServer variables, and passes them via the URL
> > string, using an http-equiv redirect. (yuck)
> >
> > I would have included the URL for the site, but the information is
> > confidential (client request...sorry).  Thanks in advance.
> >
> >
> > Cheers,
> > Jim
> >  ___________________________________    GGG G   AAA      III    AAA
> > |  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A
> > |                                   | G     G    A A      I      A A
> > |Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A
> > |Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A
> > |Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A
> > |_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA
> >            ICQ: 15567315              G     G  A     A    I    A     A
> >       http://www.gaiaenv.com           G    G  A     A    I    A     A
> >                                         GGGG  AAA   AAA  III  AAA   AAA
> >


From bowenj at gaiaenv.com  Fri Oct 29 08:49:25 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Fri, 29 Oct 1999 09:49:25 -0600
Subject: Query template file question
References: <199910291531.KAA07517@philbert.dnr.state.mn.us>
Message-ID: <3819C205.F1445DDD@gaiaenv.com>

Thanks Ken,

MapServer was reporting that my php3 file was not a valid template file
(can't remember the exact error). I will give your suggestion a try
later today (along with Brent's).

It appears you have been passing variables to files (viz., apps. and
scripts) other than plain ol' hmtl files.  Thanks for your examples.


Cheers,
Jim

kenboss at dilbert.dnr.state.mn.us wrote:
> 
> Jim--
> 
> If I'm understanding your question correctly, all you need to do is to put the URL in quotes after the
> "template" keyword in the query block in your map file.  So, instead of something like this:
> 
> LAYER
>   ...
>   ...
>   QUERY
>     TEMPLATE "http://www.dnr.state.mn.us/someapp/somefile.html"
>   END
> END
> 
> You have something like this:
> 
> LAYER
>   ...
>   ...
>   QUERY
>     TEMPLATE "http://www.dnr.state.mn.us/perl/lk_search.pl?var1=[SOMEFIELD]&var2=[ANOTHER_FIELD]"
>   END
> END
> 
> Where [SOMEFIELD] and [ANOTHER_FIELD] are attributes of the shape that the user clicked on (eg., the
> unique id you want to pass to your PHP3 script).
> 
> This can have all sorts of cool applications, for example, we use:
> 
> LAYER
>   NAME counties
>   TYPE polyline
>   STATUS default
>   DATA counties
>   CLASS
>     NAME "County Boundaries"
>     SYMBOL 1
>     SIZE 1
>     COLOR 0 0 0
>   END
>   QUERY
>     TEMPLATE "http://www.ra.dnr.state.mn.us/cgi-bin/mapserv3.2?map=photos/etc/photos.map&mapext=[shpext]"
>   END
> END # Layer counties
> 
> to allow users to zoom in to the extent of a county from an initial statewide interface (see
> http://www.ra.dnr.state.mn.us/photos/photos_init.html).  Note that in this case [shpext] is not an
> attribute of the shape per se (that is, it's not stored in the .dbf file), but is a value calculated by
> the mapserver software (documented at http://mapserver.gis.umn.edu/templates.html).
> 
> Hope this helps,
> 
> --Ken
> 
> ============================================================================
> Ken Boss
> Digital Image Analysis / Web Stuff      Forestry Resource Assessment
> kenboss at dilbert.dnr.state.mn.us         Minnesota Dept. of Natural Resources
> Voice: 218 327 4449 ext. 237            413 SE 13th Street
> Fax:   218 327 4517                     Grand Rapids, MN  55744         USA
> ============================================================================

-- 
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA


From steve.lime at dnr.state.mn.us  Fri Oct 29 11:32:56 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 29 Oct 1999 13:32:56 -0500
Subject: Cool client and utilities...
Message-ID: <s819a21c.043@smtp.dnr.state.mn.us>

Hi all: I'm sure most of you are aware of the JShape java GIS client app.
Seems like really nice software although I haven't played with it. The author
has also developed a shapefile utility called FShape for pre-processing shape
files. For example, you can tile and/or generalize a shapefile. It supports some
sort of indexing but it's not compatible with the MapServer. All this is done in
Java so I don't suppose it's the speediest of tools but it seems pretty useful
regardless of the other tools you use. Anyone tried it?

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

(651) 297-2937
steve.lime at dnr.state.mn.us



From pgallen at randomlogic.com  Fri Oct 29 18:39:04 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 29 Oct 1999 18:39:04 -0700
Subject: Buttons
Message-ID: <381A4C38.94519D85@randomlogic.com>

Is it possible to make two buttons on the the map interface to zoom in or zoom
out? So far I've only been able to make one that redraws the map (with the rest
of the web page) full size - just like it is when the map page is first drawn.

PGA



From pgallen at randomlogic.com  Fri Oct 29 20:31:15 1999
From: pgallen at randomlogic.com (Paul G. Allen)
Date: Fri, 29 Oct 1999 20:31:15 -0700
Subject: Java code...
References: <s816faf2.037@smtp.dnr.state.mn.us> <381A6401.3B90E72B@randomlogic.com>
Message-ID: <381A6683.DBD22801@randomlogic.com>

I meant this to go to the list the first time. I forgot to mention, I'm running
a RedHat Linux 6.0 system and Apache, so if this is some kind of Winsucks J++
code, please let me know 'cause it won't work for me.

Thanks,

PGA

"Paul G. Allen" wrote:

> I followed the directions, and used the dnr.state.mn.us site as a guide. I get
> an error in my browser when I try to load the page "Applet mapimage class
> mapimage could not be loaded". I tried to compile the source again, and got this
> error:
> 
> Connecting...
> javac  /home/httpd/java/mapimage.java
> error "/home/httpd/java/mapimage.java" 14  class JSObject not found in class
> mapimage
>     JSObject    win;
>     ^
> error "/home/httpd/java/mapimage.java" 29  variable JSObject not found in class
> mapimage
>         win = JSObject.getWindow(this);
>               ^
> 2 errors
> make: *** [/home/httpd/java/mapimage.class] Error 255
> Execution complete
> 
> I know absolutely nothing about Java. Can someone clue me in? I'd really like to
> use this type of interface.
> 
> Thanks,
> 
> PGA


From bowenj at gaiaenv.com  Sat Oct 30 08:59:07 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Sat, 30 Oct 1999 09:59:07 -0600
Subject: Query template file question
References: <3819B1FB.CF27D26D@gaiaenv.com>
Message-ID: <381B15CB.49F27A65@gaiaenv.com>

Hello all,

I tried both Brent and Ken's suggestions with limited success...perhaps
I am missing something simple.  Here are my findings:


**********************************************************************
*** Brent's suggestion (change the mapserv.ini setting) **************
**********************************************************************
As a simple test, I created a file with no php3 tags, simply straight
html [with a couple of MapServer tags (e.g., [shpminx])] and gave the
file an extension of '.php3'.  I changed the mapserv.ini file to reflect
the '.php3' extension for templates.  The query worked and MapServer
populated the template with information; however, if I try to redraw the
map, I receive a 'malformed template' error from MapServer.  Any
thoughts?  

Also, does anyone know if it is possible to include more that one
regular expression per parameter in mapserv.ini? (e.g.,
'TEMPLATE_EXTENSION=\.html$ or TEMPLATE_EXTENSION=\.php3$')  I tried
this, but received an error 'received 3 parameters, expecting 2'.  FYI,
I also tried 'TEMPLATE_EXTENSION=\.html$ or \.php3$'


**********************************************************************
*** Ken's suggestion (pass args directly to PHP3 via the URL str) ****
**********************************************************************
I attempted to pass the arguments to my php3 page as suggested by Ken.  

I noticed I was not double-quoting my path and filename parameter in my
map file.  The map file example
(http://mapserver.gis.umn.edu/demo_map.txt) does not indicate
double-quotes are required.  It appears the double-quotes are required
only if special characters (e.g., ?, %, etc.) are included in the
parameter (my guess).  Changing to double-quoted string permitted the
passing of arguments from MapServer to PHP3 without error.  The result
was no 'malformed template' error when drawing the map, but this error
did arise when querying.


**************
*** Other ****
**************
I also tried hybrids of Brent's and Ken's suggestions.  Does anyone have
any other ideas surrounding this issue?  

If there are no conclusions in the next couple days, I suggest killing
this thread.  The 'http-equiv refresh' workaround is not too cumbersome
and it does produce results.  Thanks in advance.


Cheers,
Jim
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA


From steve.lime at dnr.state.mn.us  Sat Oct 30 08:59:49 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Sat, 30 Oct 1999 10:59:49 -0500
Subject: Java code...
Message-ID: <s81acfb8.005@smtp.dnr.state.mn.us>

No, pure Java. Wrote latest version on a Linux box. You need to add the netscape java classes to
your CLASSPATH environment variable. Mine looks like this on RH 6.0.

 CLASSPATH=/usr/local/jdk117_v3/lib/classes.zip:/usr/lib/netscape/java/classes/java40.jar

Yours would vary depending on what JDK your using and what version of communicator is laying around.
Unless you're adding stuff to the src you shouldn't nee to recompile and can just use the class file I sent.

Steve

<<< "Paul G. Allen" <pgallen at randomlogic.com> 10/29 10:32p >>>
I meant this to go to the list the first time. I forgot to mention, I'm running
a RedHat Linux 6.0 system and Apache, so if this is some kind of Winsucks J++
code, please let me know 'cause it won't work for me.

Thanks,

PGA

"Paul G. Allen" wrote:

> I followed the directions, and used the dnr.state.mn.us site as a guide. I get
> an error in my browser when I try to load the page "Applet mapimage class
> mapimage could not be loaded". I tried to compile the source again, and got this
> error:
> 
> Connecting...
> javac  /home/httpd/java/mapimage.java
> error "/home/httpd/java/mapimage.java" 14  class JSObject not found in class
> mapimage
>     JSObject    win;
>     ^
> error "/home/httpd/java/mapimage.java" 29  variable JSObject not found in class
> mapimage
>         win = JSObject.getWindow(this);
>               ^
> 2 errors
> make: *** [/home/httpd/java/mapimage.class] Error 255
> Execution complete
> 
> I know absolutely nothing about Java. Can someone clue me in? I'd really like to
> use this type of interface.
> 
> Thanks,
> 
> PGA



From steve.lime at dnr.state.mn.us  Sat Oct 30 09:22:55 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Sat, 30 Oct 1999 11:22:55 -0500
Subject: Query template file question
Message-ID: <s81ad51c.008@smtp.dnr.state.mn.us>

Ok. The comments in the .ini file say that the value for TEMPLATE_VALUE should be a regular
expression. The examples you show are not regular
expresions, hence there problems. O'Reilly has an
excellent regex reference book in case anyone in
is interested. You have the right idea (or'ing), but the syntax is all wrong.  

Corrent me if I'm wrong , but CGI processes return
content directly to the browser. I don't believe
you can have PHP process a page generated by the
MapServer. I know this used to be the case but
haven't researched it in a while.

In general, I've chained things together using URL
forwarding. MapServer forwards its stuff to another URL, bypassing the browser. That's what Ken was suggesting. Other than the double quotes
issue it sounds as if that worked for you.

Another way to do it is to use JavaScript. You can
create a page that is essentially empty except for
some javascript. That code can be initialized using standard mapserver substitutions and then can either build the page itself or replace the page with some URL. Example 5 in the short tutorial I mention on the list in late Sept does exactly this. Can be very powerful. 

If anyone knows of a way to filter a page from CGI
before it is sent to the client, speak up. I'd love to be able to even just use simple SSI in a
template file.

Of course, with mapscript much of this becomes a
thing of the past as you can do all of your mapping and PHP-like stuff (eg. database) in one
script.

Steve

 

<<< "Jim T. Bowen (GAIA Consultants Inc.)" <bowenj at gaiaenv.com> 10/30 10:54a >>>
Hello all,

I tried both Brent and Ken's suggestions with limited success...perhaps
I am missing something simple.  Here are my findings:


**********************************************************************
*** Brent's suggestion (change the mapserv.ini setting) **************
**********************************************************************
As a simple test, I created a file with no php3 tags, simply straight
html [with a couple of MapServer tags (e.g., [shpminx])] and gave the
file an extension of '.php3'.  I changed the mapserv.ini file to reflect
the '.php3' extension for templates.  The query worked and MapServer
populated the template with information; however, if I try to redraw the
map, I receive a 'malformed template' error from MapServer.  Any
thoughts?  

Also, does anyone know if it is possible to include more that one
regular expression per parameter in mapserv.ini? (e.g.,
'TEMPLATE_EXTENSION=\.html$ or TEMPLATE_EXTENSION=\.php3$')  I tried
this, but received an error 'received 3 parameters, expecting 2'.  FYI,
I also tried 'TEMPLATE_EXTENSION=\.html$ or \.php3$'


**********************************************************************
*** Ken's suggestion (pass args directly to PHP3 via the URL str) ****
**********************************************************************
I attempted to pass the arguments to my php3 page as suggested by Ken.  

I noticed I was not double-quoting my path and filename parameter in my
map file.  The map file example
(http://mapserver.gis.umn.edu/demo_map.txt) does not indicate
double-quotes are required.  It appears the double-quotes are required
only if special characters (e.g., ?, %, etc.) are included in the
parameter (my guess).  Changing to double-quoted string permitted the
passing of arguments from MapServer to PHP3 without error.  The result
was no 'malformed template' error when drawing the map, but this error
did arise when querying.


**************
*** Other ****
**************
I also tried hybrids of Brent's and Ken's suggestions.  Does anyone have
any other ideas surrounding this issue?  

If there are no conclusions in the next couple days, I suggest killing
this thread.  The 'http-equiv refresh' workaround is not too cumbersome
and it does produce results.  Thanks in advance.


Cheers,
Jim
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA



From bowenj at gaiaenv.com  Sat Oct 30 10:40:31 1999
From: bowenj at gaiaenv.com (Jim T. Bowen (GAIA Consultants Inc.))
Date: Sat, 30 Oct 1999 11:40:31 -0600
Subject: Query template file question
References: <s81ad51c.007@smtp.dnr.state.mn.us>
Message-ID: <381B2D8F.F56ADBEB@gaiaenv.com>

Stephen Lime wrote:
> 
> Ok. The comments in the .ini file say that the value for TEMPLATE_VALUE should be a regular
> expression. The examples you show are not regular
> expresions, hence there problems. O'Reilly has an
> excellent regex reference book in case anyone in
> is interested. You have the right idea (or'ing), but the syntax is all wrong.

Thanks...too much code in my head and not enough time in the day!  I
came accross this site...may be a good intro. link for those 'still
climbing':   http://www.engr.usask.ca/~macphed/UNIX/regexp.html

> 
> Corrent me if I'm wrong , but CGI processes return
> content directly to the browser. I don't believe
> you can have PHP process a page generated by the
> MapServer. I know this used to be the case but
> haven't researched it in a while.

To the best of my knowledge, you are right re. the flow of results from
a CGI process.  This is likely one of the insurmountable issues I face,
even if I overcome the 'malformed template' issue.  It may be that
http-equiv or URL forwarding is the easiest/quickest way to address the
issue at this time.  

> 
> In general, I've chained things together using URL
> forwarding. MapServer forwards its stuff to another URL, bypassing the browser. That's what Ken was suggesting. Other than the double quotes
> issue it sounds as if that worked for you.

Not really.  When querying, MapServer still reports a 'malformed
template' when I pass the argument-based URL string to a PHP3 file.

> 
> Another way to do it is to use JavaScript. You can
> create a page that is essentially empty except for
> some javascript. That code can be initialized using standard mapserver substitutions and then
> can either build the page itself or replace the page with some URL. Example 5 in the short 
>tutorial I mention on the list in late Sept does exactly this. Can be very powerful.
> 
> If anyone knows of a way to filter a page from CGI
> before it is sent to the client, speak up. I'd love to be able to even just use simple SSI in a
> template file.

Not sure...would be challenging based on our dialog re. flow between
browser and CGI.  I too would be interested if anyone knows of a means
to accomplish this.

> 
> Of course, with mapscript much of this becomes a
> thing of the past as you can do all of your mapping and PHP-like stuff (eg. database) in one
> script.
> 

I look forward to mapscript!  :)

> Steve


Cheers,
Jim
 ___________________________________    GGG G   AAA      III    AAA     
|  Jim T. Bowen, M.Sc., President   |  G   GG     A       I       A     
|                                   | G     G    A A      I      A A    
|Suite 306,  822-11th Ave. S.W.     | G     G    A A      I      A A    
|Calgary, Alberta, Canada  T2R 0E5  | G         A   A     I     A   A   
|Ph: 403-571-7213  Fax: 403-571-7211| G   GGGG  A   A     I     A   A   
|_______bowenj at gaiaenv.com__________| G     G   AAAAA     I     AAAAA   
           ICQ: 15567315              G     G  A     A    I    A     A  
      http://www.gaiaenv.com           G    G  A     A    I    A     A
                                        GGGG  AAA   AAA  III  AAA   AAA


From marin at here.is  Sun Oct 31 06:15:09 1999
From: marin at here.is (Ma)
Date: Sun, 31 Oct 1999 14:15:09 +0000
Subject: offsite, transparent color, gd1.3
Message-ID: <381C4EED.962E39B7@here.is>


1. What is  "OFFSITE 0" in the raster layer definition (see demo.map)?
2. How to specify transparent color for the raster layer?
3. How to "make" mapserver3.2 with gd1.3?


--
Ma




From steve.lime at dnr.state.mn.us  Sun Oct 31 19:42:21 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Sun, 31 Oct 1999 21:42:21 -0600
Subject: offsite, transparent color, gd1.3
Message-ID: <s81cb7dc.010@smtp.dnr.state.mn.us>

Question 2 answers question 1. Offsite is the color index value in the source image to make transparent. "OFFSITE 0" means that all source pixels with value 0 become transparent in the output image.

To build with 1.3 you should just have to switch
the references in the Makefile to point to the other dist. I don't know if 1.3 is where the GD
lib went from column major to row major. If so there are a few other fixes that have to happen. 

GD is not used at all in processing TRUETYPE fonts
so this will not remedy the character mapping issues with latin chars. May fix that with BITMAP
fonts assuming the GD version include those characters. Haven't had time to test your sample
data yet, will do this week sometime. 

Steve

<<< Ma <marin at here.is> 10/31  8:14a >>>

1. What is  "OFFSITE 0" in the raster layer definition (see demo.map)?
2. How to specify transparent color for the raster layer?
3. How to "make" mapserver3.2 with gd1.3?


--
Ma