From steve.lime at dnr.state.mn.us  Wed Jun  9 09:07:55 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Wed, 09 Jun 1999 11:07:55 -0500
Subject: Version 3.217 available.
Message-ID: <s75e4b20.027@smtp.dnr.state.mn.us>

Greetings: MapServer version 3.217 is available at the MapServer homepage.
Besides fixing several bugs found in porting to NT it introduces the following
features:

  - Label buffering. You can know specify a buffer distance (in pixels) for leaving
    space around labels. This can greatly enhance map readability. (Note: this
    option only works with horizontal text at this point)

  - One-to-many joins. Joins have been reworked and are now a separate object
    within a query object. One-to-one and one-to-many joins are supported.

  - Polygon labeling. Polygon label points are now gauranteed to fall within the
    polygon. Previously certain polygons may have had labels fall outsite polygon
    or within islands. This has been fixed. If the polygon centroid does not fall in
    the polygon then an alternative method is used. The label point will be in the
    polygon, but text may still overlap other features.

As always, please let me know of problems, questions or whatever. Thanks.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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



From bfraser at digitalscience.com  Thu Jun 10 13:18:48 1999
From: bfraser at digitalscience.com (Brent Fraser)
Date: Thu, 10 Jun 1999 14:18:48 -0600
Subject: Version 3.217 available.
References: <s75e4b20.027@smtp.dnr.state.mn.us>
Message-ID: <006501beb37e$747a35d0$390002c0@servo>

Steve,

  I downloaded and compiled 3.217.  Looks Great!
I noticed you've added some keywords in the map file
syntax (although I guess they could have been there in
previous versions):

LABEL
    ANGLE
    PARTIALS

LEGEND:
    KEYSIZE
    KEYSPACING


And now for another request...

  I've been pursuing labeling with shapefiles for a while now,
with the goal of making a cartographically pleasing map.
As you know, shapefiles have no inherent text label object,
the software used to view the shapefile must somehow
use the geometry in the shp portion for the postion and
possibly the rotation, and get the size and other text
parameters from somewhere else.  OK for GIS, but not
good for mapping.

And its really frustrating when the original CAD files have
the text carefully placed, sized and rotated!  ArcView will
honour the size and rotation when reading CAD files (it has
CAD file support), but doesn't do a good job of saving that
information with the text value.  But there are other ways to
convert CAD to shape....

Anyways,  I'd like to promote (and have mapserv support) a
shapefile convention of using DBF fields for the ANGLE and
SIZE (and of course the TEXT value) text attributes.  This
would enable individual text strings to have their own size
and rotation, and get pretty close to a great looking map.

Your thoughts?

Thanks,
Brent




----- Original Message -----
From: Stephen Lime <steve.lime at dnr.state.mn.us>
To: <mapserver-users at lists.gis.umn.edu>
Sent: Wednesday, June 09, 1999 10:07 AM
Subject: Version 3.217 available.


> Greetings: MapServer version 3.217 is available at the MapServer homepage.
> Besides fixing several bugs found in porting to NT it introduces the
following
> features:
>
>   - Label buffering. You can know specify a buffer distance (in pixels)
for leaving
>     space around labels. This can greatly enhance map readability. (Note:
this
>     option only works with horizontal text at this point)
>
>   - One-to-many joins. Joins have been reworked and are now a separate
object
>     within a query object. One-to-one and one-to-many joins are supported.
>
>   - Polygon labeling. Polygon label points are now gauranteed to fall
within the
>     polygon. Previously certain polygons may have had labels fall outsite
polygon
>     or within islands. This has been fixed. If the polygon centroid does
not fall in
>     the polygon then an alternative method is used. The label point will
be in the
>     polygon, but text may still overlap other features.
>
> As always, please let me know of problems, questions or whatever. Thanks.
>
> 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  Thu Jun 10 13:57:32 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Thu, 10 Jun 1999 15:57:32 -0500
Subject: Version 3.217 available.
Message-ID: <s75fe085.064@smtp.dnr.state.mn.us>

Brent et al. So many options these days I forget to mention them. The legend
options have always been there. There others are relatively new, or brand new.
ANGLE specifies a text rotation angle (0-360). It does work, but it's not automatically
computed and offets from markers aren't quite right. Still working on it. The
PATIALS keywords tells the labeling routine to make sure labels don't run off the
edge of a map. Values are on or off with default being on. Available only with cached
labels and no overlap.

Item based symbol and text parameters are a good idea. Versions earlier than 3.2
supported reading symbol, color and size from dbf files. The only ones that make
sense to me anymore are symbol size, text size and angle. I'll look into adding that.
Only drag is the speed loss you must endure with all those lookups.

Other comments?

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> "Brent Fraser" <bfraser at digitalscience.com> 06/10 3:18 PM >>>
Steve,

  I downloaded and compiled 3.217.  Looks Great!
I noticed you've added some keywords in the map file
syntax (although I guess they could have been there in
previous versions):

LABEL
    ANGLE
    PARTIALS

LEGEND:
    KEYSIZE
    KEYSPACING


And now for another request...

  I've been pursuing labeling with shapefiles for a while now,
with the goal of making a cartographically pleasing map.
As you know, shapefiles have no inherent text label object,
the software used to view the shapefile must somehow
use the geometry in the shp portion for the postion and
possibly the rotation, and get the size and other text
parameters from somewhere else.  OK for GIS, but not
good for mapping.

And its really frustrating when the original CAD files have
the text carefully placed, sized and rotated!  ArcView will
honour the size and rotation when reading CAD files (it has
CAD file support), but doesn't do a good job of saving that
information with the text value.  But there are other ways to
convert CAD to shape....

Anyways,  I'd like to promote (and have mapserv support) a
shapefile convention of using DBF fields for the ANGLE and
SIZE (and of course the TEXT value) text attributes.  This
would enable individual text strings to have their own size
and rotation, and get pretty close to a great looking map.

Your thoughts?

Thanks,
Brent




----- Original Message -----
From: Stephen Lime <steve.lime at dnr.state.mn.us>
To: <mapserver-users at lists.gis.umn.edu>
Sent: Wednesday, June 09, 1999 10:07 AM
Subject: Version 3.217 available.


> Greetings: MapServer version 3.217 is available at the MapServer homepage.
> Besides fixing several bugs found in porting to NT it introduces the
following
> features:
>
>   - Label buffering. You can know specify a buffer distance (in pixels)
for leaving
>     space around labels. This can greatly enhance map readability. (Note:
this
>     option only works with horizontal text at this point)
>
>   - One-to-many joins. Joins have been reworked and are now a separate
object
>     within a query object. One-to-one and one-to-many joins are supported.
>
>   - Polygon labeling. Polygon label points are now gauranteed to fall
within the
>     polygon. Previously certain polygons may have had labels fall outsite
polygon
>     or within islands. This has been fixed. If the polygon centroid does
not fall in
>     the polygon then an alternative method is used. The label point will
be in the
>     polygon, but text may still overlap other features.
>
> As always, please let me know of problems, questions or whatever. Thanks.
>
> Steve
>
> Stephen Lime
> Internet Applications Analyst
> MIS Bureau - MN DNR
>
> (651) 297-2937
> steve.lime at dnr.state.mn.us 
>
>




From mapadm at polsl.gliwice.pl  Fri Jun 11 01:48:18 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Fri, 11 Jun 1999 10:48:18 +0200
Subject: Version 3.217 available.
In-Reply-To: <s75fe085.064@smtp.dnr.state.mn.us>
Message-ID: <199906110847.KAA20343@zeus.polsl.gliwice.pl>

Date sent:      	Thu, 10 Jun 1999 15:57:32 -0500
From:           	"Stephen Lime" <steve.lime at dnr.state.mn.us>
To:             	<bfraser at digitalscience.com>
Copies to:      	<mapserver-users at lists.gis.umn.edu>
Subject:        	Re: Version 3.217 available.

> >>> "Brent Fraser" <bfraser at digitalscience.com> 06/10 3:18 PM >>>
> >>>"Stephen Lime" <steve.lime at dnr.state.mn.us>
Hi,

There is quite easy way to add rotated label for every individual 
LINE shape.
I used your "Find Middle" procedure Steve, you are calculating the 
angle there :), added two lines and "Find Angle" is ready.

------------------------------
  /*
  ** point we want lies between point i and i-1
  */  
  if (p[i].y==p[i-1].y)
	  return(0);
  theta = asin(ms_abs(p[i].x - p[i-1].x)/sqrt((pow((p[i].x - p[i-1].x),2) 
+ pow((p[i].y - p[i-1].y),2))));

  if ( ((p[i].y < p[i-1].y) && (p[i].x < p[i-1].x)) || ((p[i].y > p[i-1].y) && 
(p[i].x > p[i-1].x))  )
	  theta = theta + 2*PI;
  return(ms_abs(theta - PI/2));
------------------------------

As a point of placement may be used CR label position (eventually 
with "advance_x" in gdttf set to negative value of half string length).

Grzegorz




From mapadm at polsl.gliwice.pl  Sun Jun 13 13:07:38 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Sun, 13 Jun 1999 22:07:38 +0200
Subject: new version
Message-ID: <000101beb5d8$62d76800$b2cb9e9d@g>

In previous versions fixed font labels positioned at CC, were centered
vertically (each word) in case of wrapping , now the whole label is
centered, but not each word (token).
Is there any way to center each token ?


I have also one suggestion to Steven:
I think that most important annotations should be listed first in the layer
classes definition to ensure they are drawn if labeloverlap is false.

and one bug:
there is no min/maxscale support for labels in new version ?

Grzegorz



From steve.lime at dnr.state.mn.us  Sun Jun 13 20:57:59 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Sun, 13 Jun 1999 22:57:59 -0500
Subject: new version
Message-ID: <s764a9c6.010@smtp.dnr.state.mn.us>

Hi Grzegorz:

1- Nope, no way (currently) to center each string. That's the price for the antialiasing and font roatation in the gdft package. That code is all character based where my old stuff working on strings/tokens. It may be possible to add justification again, but it won't be easy.

2- I went back and forth on "last-in-first-out"
or "first-in-first-out". The current setup is LIFO
for a couple of reasons. First, things drawn last are presumably the most important. You must consider the label cache as spanning layers and not just features. Second, it makes sense when drawing/symbolizing roads layers, especially if different classes of roads share arcs. More important roads can be drawn after (and on top of) lesser roads, and then they can be labeled that way as well. It turns out that sorting shapefiles can be real usefull for labeling. There is a utility in the shapelib directory called shpsort (or sortshp) that will sort a shapefile based on a numerical item. I may try to do something like LIFO for layers and FIFO for features but haven't had time.

3- No, it's there and called "LABELMINSCALE" and "LABELMAXSCALE".

Steve

<<< "Grzegorz Myrda" <mapadm at polsl.gliwice.pl>  6/13  3:07p >>>
In previous versions fixed font labels positioned at CC, were centered
vertically (each word) in case of wrapping , now the whole label is
centered, but not each word (token).
Is there any way to center each token ?


I have also one suggestion to Steven:
I think that most important annotations should be listed first in the layer
classes definition to ensure they are drawn if labeloverlap is false.

and one bug:
there is no min/maxscale support for labels in new version ?

Grzegorz




From mapadm at polsl.gliwice.pl  Tue Jun 15 01:28:09 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Tue, 15 Jun 1999 10:28:09 +0200
Subject: new version
In-Reply-To: <s764a9c6.010@smtp.dnr.state.mn.us>
Message-ID: <000001beb709$00dfacc0$5510a0d4@g>

Thanks for such a detailed answer.

>
> 3- No, it's there and called "LABELMINSCALE" and "LABELMAXSCALE".
>
Oh, sorry. I got accustomed in previous version it was LABEL property, not
LAYER. So I thought I can achieve effect of disappearing smaller cities
class labels while scaling cities layer.

Thanks,
Grzegorz



From steve.lime at dnr.state.mn.us  Tue Jun 15 07:32:23 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 15 Jun 1999 09:32:23 -0500
Subject: new version
Message-ID: <s7661dc7.001@smtp.dnr.state.mn.us>

Actually you can achieve that effect. You do this using sorted shapefiles 
drawn as annotation layers. Assuming you have some variable that allows 
you to classify cities (such as population), you could sort the shapefile (using
sortshp/shpsort) on that item so that the largest cities get drawn last and
hence labeled first (remember LIFO). Since with annotation layers markers
are only drawn if labels are drawn, at small scales most small cities will overlap
with larger city names and not be drawn. As you zoom in there is more space 
for labels and gradually the smaller cities will appear, but larger cities will always
take priority. I generally like this method better than the class based scaling as
it handles areas with either very dense or very sparse feature occurences better.

There is another layer option called MAXFEATURES that limits the number of
features drawn in any one map on a layer by layer basis. In can be used to create
a similar effect. There are some problems with it and label overlap though that
I'm still working on so you get the desired results.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> "Grzegorz Myrda" <mapadm at polsl.gliwice.pl> 06/15 3:28 AM >>>
Thanks for such a detailed answer.

>
> 3- No, it's there and called "LABELMINSCALE" and "LABELMAXSCALE".
>
Oh, sorry. I got accustomed in previous version it was LABEL property, not
LAYER. So I thought I can achieve effect of disappearing smaller cities
class labels while scaling cities layer.

Thanks,
Grzegorz




From steve.lime at dnr.state.mn.us  Fri Jun 18 12:41:48 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Fri, 18 Jun 1999 14:41:48 -0500
Subject: Version 3.218b available
Message-ID: <s76a5acb.040@smtp.dnr.state.mn.us>

Howdy: Another maintenance release is available. Not many new features but
several documented features now work correctly and some bugs have been 
fixed. Here are the main changes:

 - fixed problems placing bitmapped text
 - fixed multiline bitmapped text rendering problem
 - MAXFEATURES now works correctly with label caching and label overlap removal

(Actually this turns out to be real useful. Included in the distribution is a routine
to sort shapefiles based on numeric fields. It's called sortshp and it's in the shapelib
directory in the distribution. Remember features are labeled using a "last-in-first-out"
scheme. So, you can now sort shapefiles based on area or population (ascending
order) and use MAXFEATURES to only show the biggest or most populus features
ensuring that the largest also get labeled first. I use this with lakes data, showing
only the 100 largest in each view. At small scales this makes sense, no reason to
show every 5 acre lake at a statewide scale. Definite performance increases as
well, but again it only makes sense with sorted data.)

  - fixed bug in selection of polygons with multiple outter rings and islands (faster too)
  - [shpminx] etc. are now projected if necessary
  - empty CGI variables (i.e. title=&name=test+server) are ignored

I'll be adding support for getting label size and angle from .dbf items and support for
dynamically changing layer ordering in the next few days. As always, suggestions
are welcome.

Steve


Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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



From camerons at cat.org.au  Wed Jun 23 04:03:21 1999
From: camerons at cat.org.au (Cameron Shorter)
Date: Wed, 23 Jun 1999 21:03:21 +1000
Subject: New User Questions
Message-ID: <199906231107.VAA14075@warp.ipacific.net.au>

Hello mapserver-users,

I've just subscribed to this list, so am not sure of the protocols yet.
I'm installing mapserver with the intent of providing bicycle maps for
people in Sydney/Australia.

So far I have a number of questions:
1. Is this email list achived?
2. Is there any version control of the software and demos (like CVS) and if
so, is it possible to get read access to it so I can determine if a bug has
been fixed before reporting it.
3. So far I have installed the demo (version 3.217) and I get the following
error message when I press the "Initialize" button on
http://www.cat.org.au/mapserv/ms_demo/demo_init.html :

An error has occurred in loadJoin(): Unknown identifier.
(data/lakejoin.dbf):(117) 

I've found that I can fix the problem by commenting out QUERYITEM from
demo.map : 

#  QUERYITEM usclass
#  QUERY
#    EXPRESSION "421"
#    TEMPLATE lakes.html
#    JOIN "data/lakejoin.dbf" islake "type"
#  END
#  QUERY
#  EXPRESSION "^-"
#    TEMPLATE land.html
#    JOIN "data/lakejoin.dbf" islake "type"
#  END

You can see this working at
http://www.cat.org.au/mapserv/ms_demo2/demo_init.html .
Of course the queries don't work.  So my question is how do I get the
queries to work?

Cameron Shorter
61.2.9901-1596 work.


From mapadm at polsl.gliwice.pl  Wed Jun 23 06:22:43 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Wed, 23 Jun 1999 15:22:43 +0200
Subject: New User Questions
In-Reply-To: <199906231107.VAA14075@warp.ipacific.net.au>
Message-ID: <199906231320.PAA07816@zeus.polsl.gliwice.pl>

> An error has occurred in loadJoin(): Unknown identifier.
> (data/lakejoin.dbf):(117) 
> 
> #    JOIN "data/lakejoin.dbf" islake "type"

I use full path, with no quotation-marks, field names in uppercase, 
and it works.

Grzegorz


From steve.lime at dnr.state.mn.us  Wed Jun 23 08:04:34 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Wed, 23 Jun 1999 10:04:34 -0500
Subject: New User Questions
Message-ID: <s770b137.001@smtp.dnr.state.mn.us>

Joins changed with version 3.217. There is now a join object instead of a single line
join statement. This was needed in order add support for one-to-many joins. The syntax
is similar to any other mapserver object. (it is documented on the web site) The demo
breaks because I haven't updated it yet- sorry. Anyway, here's how this would be re-written.

JOIN "data/lakejoin.dbf" islake "type"

becomes

JOIN
  NAME lakejoin
  TABLE data/lakejoin.dbf
  FROM islake # item names are case insensitive
  TO "type"  # must be quoted as TYPE is a mapserver keyword
END

I believe TABLE needs to be full path. An oversight on my part and it will be fixed in the
next release. Joined items can be accessed in templates by surrounding their name with
[]'s (eg. [TYPE]). Non-unique names can be accessed by prefacing the item name with
the join name (eg. [lakejoin_TYPE]). Note that the item names in templates are case
sensitive. ArcView always writes xbase item names in all-caps. 

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> "Grzegorz Myrda" <mapadm at polsl.gliwice.pl> 06/23 8:22 AM >>>
> An error has occurred in loadJoin(): Unknown identifier.
> (data/lakejoin.dbf):(117) 
> 
> #    JOIN "data/lakejoin.dbf" islake "type"

I use full path, with no quotation-marks, field names in uppercase, 
and it works.

Grzegorz



From steve.lime at dnr.state.mn.us  Wed Jun 23 08:12:44 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Wed, 23 Jun 1999 10:12:44 -0500
Subject: New User Questions
Message-ID: <s770b320.054@smtp.dnr.state.mn.us>

Hi: Just sent an explanation of the JOIN problem. In short comment
that line out. You'll also need to comment out a line further down
beginning with JUSTIFY. In the future I'll make sure the demo and the
release version are better synced. Queries will work without the join.

FYI 3.218 is the most recent release. Next version will be 3.3 and will
be out in a couple of weeks with some significant improvements.

Email list is archived (I think) although there is no web archive available
yet. UMN guys are working on it though. You can get old messages using
the standard majordomo commands. For a complete list send a message
with the word "help" and nothing else to majordomo at lists.gis.umn.edu.

CVS is in use at home on my development machine, but is not otherwise
accessible. This may change at some point. Bug reports are sparse, so
please let me know what you find.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> Cameron Shorter <camerons at cat.org.au> 06/23 6:03 AM >>>
Hello mapserver-users,

I've just subscribed to this list, so am not sure of the protocols yet.
I'm installing mapserver with the intent of providing bicycle maps for
people in Sydney/Australia.

So far I have a number of questions:
1. Is this email list achived?
2. Is there any version control of the software and demos (like CVS) and if
so, is it possible to get read access to it so I can determine if a bug has
been fixed before reporting it.
3. So far I have installed the demo (version 3.217) and I get the following
error message when I press the "Initialize" button on
http://www.cat.org.au/mapserv/ms_demo/demo_init.html :

An error has occurred in loadJoin(): Unknown identifier.
(data/lakejoin.dbf):(117) 

I've found that I can fix the problem by commenting out QUERYITEM from
demo.map : 

#  QUERYITEM usclass
#  QUERY
#    EXPRESSION "421"
#    TEMPLATE lakes.html
#    JOIN "data/lakejoin.dbf" islake "type"
#  END
#  QUERY
#  EXPRESSION "^-"
#    TEMPLATE land.html
#    JOIN "data/lakejoin.dbf" islake "type"
#  END

You can see this working at
http://www.cat.org.au/mapserv/ms_demo2/demo_init.html .
Of course the queries don't work.  So my question is how do I get the
queries to work?

Cameron Shorter
61.2.9901-1596 work.



From steve.lime at dnr.state.mn.us  Thu Jun 24 21:29:24 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Thu, 24 Jun 1999 23:29:24 -0500
Subject: Version 3.219
Message-ID: <s772bf65.070@smtp.dnr.state.mn.us>

Howdy: Just a maintenence release to fix a few bonehead mistakes in 3.218. A non-existant mapfile will not crash the program. Compiling with Proj.4 support will work. A debugging line of code that generates bounding boxes around text has been removed.

This will be the last release in 3.2. 3.3 will be the next release in a few weeks. (unless someone finds some other mistakes on my part)

The demo has been fixed so that it works cleanly with the latest version. A full re-work will debut with version 3.3.

Steve  


From teb at mallit.fr.umn.edu  Mon Jun 28 06:44:12 1999
From: teb at mallit.fr.umn.edu (Thomas E. Burk)
Date: Mon, 28 Jun 1999 08:44:12 -0500 (CDT)
Subject: Web interface to archives
Message-ID: <199906281344.IAA09767@mallit.fr.umn.edu>

We have put in place a _simple_ Web interface to the maapserver-users message archive. See the Support page on the mapserver Web site 
<http://mapserver.gis.umn.edu>. The interface allows access to messages by 
month, excluding the current month (which, since the current month is 06 1999, 
could be grabbed by sending a message to majordomo at lists.gis.umn.edu with the 
body {get mapserver-users mapserver-users.9906} without the braces). If list 
traffic increases we will make this more sophisticated and add searching.

Thanks,
Tom
=======================
Thomas E. Burk
Professor
University of Minnesota
tburk at forestry.umn.edu


From steve.lime at dnr.state.mn.us  Mon Jun 28 08:40:49 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Mon, 28 Jun 1999 10:40:49 -0500
Subject: Version 3.3 wish list...
Message-ID: <s777513d.021@smtp.dnr.state.mn.us>

Hi: I'm working on 3.3 and wanted to let people know what to expect featurewise
and to ask again for bug reports and feature requests. I've already gotten a number
of them and will incorporate as many as possible. Here's a list of new stuff or changes.

 - label overlap removal will now be the normal mode of operation, cached labels cannot overlap
     (labeloverlap option goes away...)
 - support for label angle and size from shapefile attributes
 - simple multi-item query support
 - addition of tileitem as means of supporting tiling by filename in addition to tiling by directory
 - selected features graphics (eg. show selected features in a hi-lite color)
 - layer ordering control
 - text underlines
 - the ability to define features (point,line,polygon) directly in the mapfile and via URL's
 - fully functional label rotation w/buffering
 - fixed legend creation with multiline labels
 - query of tiled data
 - raster data classes (i.e. reclass and recolor on-the-fly)
 - true type fonts for scalebars
 - additional scalebar styles
 - named symbols (eg. SYMBOL cross or SYMBOL circle)

and the most important addition, MapScript. MapScript is a Perl interface to the MapServer
C API. All relevant C functions and internal data structures will be accessible using Perl. This
means users will be able to write custom online or offline map creation scripts pretty easily. With
Perl's database and network connectivity this means you'll be able to pull features from Oracle
or other remote data sources (eg. text files) and add them to a map. Users could control colors,
symbols and expressions. Create graphics using on module and embed them into the map. Lots
of possiblities.

Again, now is the time to make suggestions, feature requests and bug reports. Thanks.

Steve


Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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



From kenboss at dilbert.dnr.state.mn.us  Mon Jun 28 09:07:46 1999
From: kenboss at dilbert.dnr.state.mn.us (kenboss at dilbert.dnr.state.mn.us)
Date: Mon, 28 Jun 1999 11:07:46 -0500 (CDT)
Subject: Version 3.3 wish list...
Message-ID: <199906281607.LAA25564@philbert.dnr.state.mn.us>

Steve--

How about mapfile-based definition of classes for thematic rasters, following 
the model of the shapefile stuff.  It'd be really cool to be able to do 
hierarchical class collapsing this way, rather than having to maintain separate 
rasters for that purpose.  With the advent of MapScript, we'd then have the 
capability to let users define through html forms how they want to see the 
classes displayed.  As a side benefit, we'd also get dynamic legend creation for 
raster images out of the deal.  Of course we'd also want to retain the current 
mechanisms as a default mode, so that if no classes are defined in the mapfile 
for a raster, then it is displayed in accordance with the associated color 
table.  I realize that this may be a rather large pain to develop, given the 
varying requirements for the different raster formats supported, but as long as 
we're talking wish lists, I thought I'd throw that in...

--Ken


From camerons at cat.org.au  Mon Jun 28 13:58:31 1999
From: camerons at cat.org.au (Cameron Shorter)
Date: Tue, 29 Jun 1999 06:58:31 +1000
Subject: Version 3.3 wish list...
In-Reply-To: <s777513d.021@smtp.dnr.state.mn.us>
Message-ID: <m10yiYW-000YPTC@mailarray.mpx.com.au>

* How about a version option, so you can query a mapserv package and
determine what version it is running.

* I'm also want an operator on the host to be able to enter a map feature,
post it to mapserv, and for mapserv to add it into an existing map on the
server.  This probably involves using Java on the host machine.  I'm
prepared to help code this.
Cameron Shorter
61.2.9901-1596 work.


From TBanister at SpaceImaging.com  Mon Jun 28 14:34:14 1999
From: TBanister at SpaceImaging.com (Banister Trevor)
Date: Mon, 28 Jun 1999 15:34:14 -0600
Subject: Version 3.3 wish list...
Message-ID: <AEEF94298C10D311B5A000A024B359E102D98E@imager9.spaceimaging.com>

I like this idea of more control from forms. I'm having a tough time getting
granular control without building lots of .map files.

How about an intermediate step which allows all the parameters from a .map
file to be called from a form, or from the query string? This technique
would help in using mapserver as a backend workhorse to serve images to
multiple front end servers.

This would also allow other front end systems (running PHP, Cold Fusion,
ASP, etc.) access to generate layer parameters without having to write a
temporary .map file on the mapserver, on the fly (a kludgey technique I'm
using to change the order and projection in which layers are presented)

I'm toying with the idea of letting users upload their own layers (shapes or
images) to add value to the maps the generate. Being able to generate a map
without a .map file would make this much easier too.



Since this is my first posting, let me take a minute to say thanks for
creating mapserv, and making it open source. I hope to be able to contribute
some code once I get up to speed on its internals.

Does anyone else have an interest in serving up MrSID images (with sid world
files)?
For those that don't already know, MrSID is a wavelet compression system
developed by Lizard tech (www.lizardtech.com) which is lossy like JPEG, but
produces much better compression (20:1) with hardly any loss.


Trevor Banister
Sr. Application Designer (and builder, deployer, maintainer, .....)
Space Imaging
TBanister at SpaceImaging.com

To satisfy the legal department:
The preceding statements and opinions are my own and not those of my
employer, Space Imaging.

-----Original Message-----
From: kenboss at dilbert.dnr.state.mn.us
[mailto:kenboss at dilbert.dnr.state.mn.us]
Sent: Monday, June 28, 1999 10:08 AM
To: steve.lime at dnr.state.mn.us; mapserver-users at lists.gis.umn.edu
Subject: Re: Version 3.3 wish list...


Steve--

How about mapfile-based definition of classes for thematic rasters,
following 
the model of the shapefile stuff.  It'd be really cool to be able to do 
hierarchical class collapsing this way, rather than having to maintain
separate 
rasters for that purpose.  With the advent of MapScript, we'd then have the 
capability to let users define through html forms how they want to see the 
classes displayed.  As a side benefit, we'd also get dynamic legend creation
for 
raster images out of the deal.  Of course we'd also want to retain the
current 
mechanisms as a default mode, so that if no classes are defined in the
mapfile 
for a raster, then it is displayed in accordance with the associated color 
table.  I realize that this may be a rather large pain to develop, given the

varying requirements for the different raster formats supported, but as long
as 
we're talking wish lists, I thought I'd throw that in...

--Ken


From mapadm at polsl.gliwice.pl  Tue Jun 29 03:51:36 1999
From: mapadm at polsl.gliwice.pl (Grzegorz Myrda)
Date: Tue, 29 Jun 1999 12:51:36 +0200
Subject: Version 3.3 wish list...
In-Reply-To: <s777513d.021@smtp.dnr.state.mn.us>
Message-ID: <199906291049.MAA16502@zeus.polsl.gliwice.pl>

Hi,
I think adding features on-line as suggested by Cameron can be 
done without Java, only for point objects. After selecting 
mode=ADD and clicking on the map, new feature would be added 
to the specified layer.
Thanks to avoiding Java it would be still low-requirements system.
However adding line and polygon features would really need Java 
and much of work.

And what about line features labels rotated by angle calculated 
from line angle ?

Also one Map object parameter may be added to mapfile, called 
Description or somehow, to display any fixed text at the bootom of 
the map picture (map copyrights or something).

Grzegorz


From SullivanJ at nima.mil  Mon Jun 28 10:43:21 1999
From: SullivanJ at nima.mil (Sullivan, James R.)
Date: Mon, 28 Jun 1999 13:43:21 -0400
Subject: Version 3.3 wish list...
Message-ID: <29E2BCE114FBD1118EE000805FBEA4135FDE89@WNYX2>

Has anyone thought of an avenue script for Arcview that would take the
current properties of the view to create the MapServer configuration files? 

	-----Original Message-----
	From:	kenboss at dilbert.dnr.state.mn.us
[SMTP:kenboss at dilbert.dnr.state.mn.us]
	Sent:	Monday, June 28, 1999 12:08 PM
	To:	steve.lime at dnr.state.mn.us;
mapserver-users at lists.gis.umn.edu
	Subject:	Re: Version 3.3 wish list...

	Steve--

	How about mapfile-based definition of classes for thematic rasters,
following 
	the model of the shapefile stuff.  It'd be really cool to be able to
do 
	hierarchical class collapsing this way, rather than having to
maintain separate 
	rasters for that purpose.  With the advent of MapScript, we'd then
have the 
	capability to let users define through html forms how they want to
see the 
	classes displayed.  As a side benefit, we'd also get dynamic legend
creation for 
	raster images out of the deal.  Of course we'd also want to retain
the current 
	mechanisms as a default mode, so that if no classes are defined in
the mapfile 
	for a raster, then it is displayed in accordance with the associated
color 
	table.  I realize that this may be a rather large pain to develop,
given the 
	varying requirements for the different raster formats supported, but
as long as 
	we're talking wish lists, I thought I'd throw that in...

	--Ken


From camerons at cat.org.au  Tue Jun 29 04:41:21 1999
From: camerons at cat.org.au (Cameron Shorter)
Date: Tue, 29 Jun 1999 21:41:21 +1000
Subject: Version 3.3 wish list...
In-Reply-To: <199906291049.MAA16502@zeus.polsl.gliwice.pl>
References: <s777513d.021@smtp.dnr.state.mn.us>
Message-ID: <199906291146.VAA19538@warp.ipacific.net.au>

>I think adding features on-line as suggested by Cameron can be 
>done without Java, only for point objects.

Cool, however I need an operator to enter roads, which I belive will need
lines rather than points.

Cameron Shorter
61.2.9901-1596 work.


From steve.lime at dnr.state.mn.us  Tue Jun 29 10:52:22 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 29 Jun 1999 12:52:22 -0500
Subject: Version 3.3 wish list...
Message-ID: <s778c1af.075@smtp.dnr.state.mn.us>



Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> Cameron Shorter <camerons at cat.org.au> 06/28 3:58 PM >>>
>* How about a version option, so you can query a mapserv package and
>determine what version it is running.

A comment is written to the top of each page the mapserver produces that
does give the version. I'll add code that produces this when no query is sent
as well. (eg. http://.../cgi-bin/mapserv with no params)

>* I'm also want an operator on the host to be able to enter a map feature,
>post it to mapserv, and for mapserv to add it into an existing map on the
>server.  This probably involves using Java on the host machine.  I'm
>prepared to help code this.

I'm all for this, but would prefer so build this as a separate application. MapServer
would still handle the map drawing but answering the add, delete and edit requests
would be handled by a second specialty app. Java would be necessary for coordinate
capture. Client would be the tricky part, code to do the deletes, adds etc. already
exixsts (minus the file locking etc.).

>Cameron Shorter
>61.2.9901-1596 work.

Steve



From steve.lime at dnr.state.mn.us  Tue Jun 29 11:00:18 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 29 Jun 1999 13:00:18 -0500
Subject: Version 3.3 wish list...
Message-ID: <s778c37c.049@smtp.dnr.state.mn.us>

Rotated labels for lines is on the todo. Real basic support is pretty
trivial. As pointed out in earlier messages to the list, the angle is
already be calculated.

The custom annotation you mention can already be done. You can
create layers that contain image coordinates rather than map coordinates.
There is a option called TRANSFORM, that when set to FALSE tells
the code not to convert the coordinates to image coordinates but rather
to leave them alone. This results in features that always appear in the
same spot in the map. You must create the shapefile (this will be real
easy in 3.3 with embedded features) yourself. I use the scripts in the
shapelib directory to do this. Then do something like:

LAYER
  NAME copyright
  TYPE ANNOTATION
  DATA cpyrght
  CLASS
    COLOR -1 -1 -1 # dummy color
    LABEL
      ...
    END
    TEXT "Copyright 1999 University of Minnesota"
  END
END

Remember the graphics file origin is the upper lefthand corner while map
coordinate systems use the lower lefthand corner.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> "Grzegorz Myrda" <mapadm at polsl.gliwice.pl> 06/29 5:51 AM >>>
Hi,
I think adding features on-line as suggested by Cameron can be 
done without Java, only for point objects. After selecting 
mode=ADD and clicking on the map, new feature would be added 
to the specified layer.
Thanks to avoiding Java it would be still low-requirements system.
However adding line and polygon features would really need Java 
and much of work.

And what about line features labels rotated by angle calculated 
from line angle ?

Also one Map object parameter may be added to mapfile, called 
Description or somehow, to display any fixed text at the bootom of 
the map picture (map copyrights or something).

Grzegorz



From steve.lime at dnr.state.mn.us  Tue Jun 29 12:40:59 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Tue, 29 Jun 1999 14:40:59 -0500
Subject: Fwd: Re: Version 3.3 wish list...
Message-ID: <s778db15.004@smtp.dnr.state.mn.us>

Yeah, what Chris says...
-------------- next part --------------
An embedded message was scrubbed...
From: imap at chesapeake.net
Subject: Re: Version 3.3 wish list...
Date: Tue, 29 Jun 1999 14:33:32 -0400
Size: 2872
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/19990629/2d081d41/attachment.eml>

From cshorte2 at csc.com  Tue Jun 29 16:29:39 1999
From: cshorte2 at csc.com (cshorte2 at csc.com)
Date: Wed, 30 Jun 1999 09:29:39 +1000
Subject: Version 3.3 wish list...
Message-ID: <4A25679F.00806F50.00@csc.com>

>>* I'm also want an operator on the host to be able to enter a map
feature,
>>post it to mapserv, and for mapserv to add it into an existing map on the
>>server.  This probably involves using Java on the host machine.  I'm
>>prepared to help code this.
>
>I'm all for this, but would prefer so build this as a separate
application. MapServer
>would still handle the map drawing but answering the add, delete and edit
requests
>would be handled by a second specialty app. Java would be necessary for
coordinate
>capture. Client would be the tricky part, code to do the deletes, adds
etc. already
>exixsts (minus the file locking etc.).
Most of the client code already exists.  Check out http://openmap.bbn.com/.
They have written java beans for handling mapping data.  All that I belive
is missing is the linking of the client and server, and then we would have
an end-to-end solution.

Cameron.




From rob at socialchange.net  Tue Jun 29 10:10:10 1999
From: rob at socialchange.net (Rob Atkinson)
Date: Wed, 30 Jun 1999 03:10:10 +1000
Subject: Version 3.3 wish list...
References: <AEEF94298C10D311B5A000A024B359E102D98E@imager9.spaceimaging.com>
Message-ID: <3778FDF2.34842913@socialchange.net>


To do this you will need a more extensible syntax for object specification in
the URL.
You may find this useful:

http://icmiss.socialchange.net.au/support/csgi/csgihttp.doc

The actual content is still a bit lightweight for your needs - it is more a
minimal set for interoperability, but the mechanism is extensible.

We are currently working on such issues with the OpenGIS consortium, but I'm not
at liberty to divulge details.

Also I guess it would be nice to see the map file expressed in  XML to get
instant access to parsers and validators.

Regards
Rob Atkinson


Banister Trevor wrote:

> I like this idea of more control from forms. I'm having a tough time getting
> granular control without building lots of .map files.
>
> How about an intermediate step which allows all the parameters from a .map
> file to be called from a form, or from the query string? This technique
> would help in using mapserver as a backend workhorse to serve images to
> multiple front end servers.
>
> This would also allow other front end systems (running PHP, Cold Fusion,
> ASP, etc.) access to generate layer parameters without having to write a
> temporary .map file on the mapserver, on the fly (a kludgey technique I'm
> using to change the order and projection in which layers are presented)
>
> I'm toying with the idea of letting users upload their own layers (shapes or
> images) to add value to the maps the generate. Being able to generate a map
> without a .map file would make this much easier too.
>
> Since this is my first posting, let me take a minute to say thanks for
> creating mapserv, and making it open source. I hope to be able to contribute
> some code once I get up to speed on its internals.
>
> Does anyone else have an interest in serving up MrSID images (with sid world
> files)?
> For those that don't already know, MrSID is a wavelet compression system
> developed by Lizard tech (www.lizardtech.com) which is lossy like JPEG, but
> produces much better compression (20:1) with hardly any loss.
>
> Trevor Banister
> Sr. Application Designer (and builder, deployer, maintainer, .....)
> Space Imaging
> TBanister at SpaceImaging.com
>
> To satisfy the legal department:
> The preceding statements and opinions are my own and not those of my
> employer, Space Imaging.
>
> -----Original Message-----
> From: kenboss at dilbert.dnr.state.mn.us
> [mailto:kenboss at dilbert.dnr.state.mn.us]
> Sent: Monday, June 28, 1999 10:08 AM
> To: steve.lime at dnr.state.mn.us; mapserver-users at lists.gis.umn.edu
> Subject: Re: Version 3.3 wish list...
>
> Steve--
>
> How about mapfile-based definition of classes for thematic rasters,
> following
> the model of the shapefile stuff.  It'd be really cool to be able to do
> hierarchical class collapsing this way, rather than having to maintain
> separate
> rasters for that purpose.  With the advent of MapScript, we'd then have the
> capability to let users define through html forms how they want to see the
> classes displayed.  As a side benefit, we'd also get dynamic legend creation
> for
> raster images out of the deal.  Of course we'd also want to retain the
> current
> mechanisms as a default mode, so that if no classes are defined in the
> mapfile
> for a raster, then it is displayed in accordance with the associated color
> table.  I realize that this may be a rather large pain to develop, given the
>
> varying requirements for the different raster formats supported, but as long
> as
> we're talking wish lists, I thought I'd throw that in...
>
> --Ken

--
--------------------------------------------------------------------------------

Rob Atkinson                                                rob at socialchange.net

Director                                                        Mob 0419 202 973

Social Change Online Pty Ltd                                Ph  +61 (2)9557 6500

http://online.socialchange.net                              Fax +61 (2)9519 8940





From steve.lime at dnr.state.mn.us  Wed Jun 30 09:16:50 1999
From: steve.lime at dnr.state.mn.us (Stephen Lime)
Date: Wed, 30 Jun 1999 11:16:50 -0500
Subject: Version 3.3 wish list...
Message-ID: <s779fcc9.077@smtp.dnr.state.mn.us>

I'm pondering a url based spec using an object syntax common to a number of
OOP languages, JavaScript comes to mind as a good example. Each element
in the mapfile could be refered to using simple object.member notation. So,
changing a layer classes label size would be something like:

  map.covertype.forest.label.size=5

where covertype is a unique layername and forest is a class name. Anything, from
scalebar parameters to projection arguments (a bit harder, you'd need array notation)
could be controlled this way. I just need to figure out how to use to lexers in the
same package. This would have limited use with get requests, but I suspect a few
changes would all that would happen at anyone time. The whole mapfile could be
handled using a post. I've already added the concept of layer groups to the code
so unique layer names won't take away any of the neat things you can do now. I
really like this idea.

Again, Rob fill me in on the OpenGIS stuff when you can. I hope the MapServer
group will be participating in the OGIS web technology testbed at somepoint.

The XML idea has merit. The more standards based the better. Actually the format
is REAL close, it's all tag based now. just replace LAYER with <LAYER> and the 
corresponding END with </LAYER> and so on.

Steve

Stephen Lime
Internet Applications Analyst
MIS Bureau - MN DNR

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

>>> Rob Atkinson <rob at socialchange.net> 06/29 12:10 PM >>>

To do this you will need a more extensible syntax for object specification in
the URL.
You may find this useful:

http://icmiss.socialchange.net.au/support/csgi/csgihttp.doc 

The actual content is still a bit lightweight for your needs - it is more a
minimal set for interoperability, but the mechanism is extensible.

We are currently working on such issues with the OpenGIS consortium, but I'm not
at liberty to divulge details.

Also I guess it would be nice to see the map file expressed in  XML to get
instant access to parsers and validators.

Regards
Rob Atkinson


Banister Trevor wrote:

> I like this idea of more control from forms. I'm having a tough time getting
> granular control without building lots of .map files.
>
> How about an intermediate step which allows all the parameters from a .map
> file to be called from a form, or from the query string? This technique
> would help in using mapserver as a backend workhorse to serve images to
> multiple front end servers.
>
> This would also allow other front end systems (running PHP, Cold Fusion,
> ASP, etc.) access to generate layer parameters without having to write a
> temporary .map file on the mapserver, on the fly (a kludgey technique I'm
> using to change the order and projection in which layers are presented)
>
> I'm toying with the idea of letting users upload their own layers (shapes or
> images) to add value to the maps the generate. Being able to generate a map
> without a .map file would make this much easier too.
>
> Since this is my first posting, let me take a minute to say thanks for
> creating mapserv, and making it open source. I hope to be able to contribute
> some code once I get up to speed on its internals.
>
> Does anyone else have an interest in serving up MrSID images (with sid world
> files)?
> For those that don't already know, MrSID is a wavelet compression system
> developed by Lizard tech (www.lizardtech.com) which is lossy like JPEG, but
> produces much better compression (20:1) with hardly any loss.
>
> Trevor Banister
> Sr. Application Designer (and builder, deployer, maintainer, .....)
> Space Imaging
> TBanister at SpaceImaging.com 
>
> To satisfy the legal department:
> The preceding statements and opinions are my own and not those of my
> employer, Space Imaging.
>
> -----Original Message-----
> From: kenboss at dilbert.dnr.state.mn.us 
> [mailto:kenboss at dilbert.dnr.state.mn.us] 
> Sent: Monday, June 28, 1999 10:08 AM
> To: steve.lime at dnr.state.mn.us; mapserver-users at lists.gis.umn.edu 
> Subject: Re: Version 3.3 wish list...
>
> Steve--
>
> How about mapfile-based definition of classes for thematic rasters,
> following
> the model of the shapefile stuff.  It'd be really cool to be able to do
> hierarchical class collapsing this way, rather than having to maintain
> separate
> rasters for that purpose.  With the advent of MapScript, we'd then have the
> capability to let users define through html forms how they want to see the
> classes displayed.  As a side benefit, we'd also get dynamic legend creation
> for
> raster images out of the deal.  Of course we'd also want to retain the
> current
> mechanisms as a default mode, so that if no classes are defined in the
> mapfile
> for a raster, then it is displayed in accordance with the associated color
> table.  I realize that this may be a rather large pain to develop, given the
>
> varying requirements for the different raster formats supported, but as long
> as
> we're talking wish lists, I thought I'd throw that in...
>
> --Ken

--
--------------------------------------------------------------------------------

Rob Atkinson                                                rob at socialchange.net 

Director                                                        Mob 0419 202 973

Social Change Online Pty Ltd                                Ph  +61 (2)9557 6500

http://online.socialchange.net                              Fax +61 (2)9519 8940