[mapserver-dev] Implementation of the msCalculateScale

[Lime, Steve D (DNR)]

I put the graphic together just to illustrate the differences between the extents GDAL reports for an image and what WMS uses.

There may be another explanation for the error you see. MapServer, the core any way, uses square pixels. So when an extent
and image dimensions are presented to the software and new extent is computed based on this assumption. This could alter
scale computations a touch depending on how you're computing things. This computed extent is guaranteed to contain the requested
extent plus a bit more perhaps depending on the aspect ratios of the requested extent and image dimensions.

Steve

From: mapserver-dev-bounces at lists.osgeo.org [mailto:mapserver-dev-bounces at lists.osgeo.org] On Behalf Of Mohamed Saâd HESSANE
Sent: Thursday, February 17, 2011 10:34 AM
To: Stephen Woodbridge
Cc: mapserver-dev at lists.osgeo.org
Subject: Re: [mapserver-dev] Implementation of the msCalculateScale

@Steve Lime : Thank you for the graphic. So if i understand, the map is generated with GDAL, which use the big extent, and the scale and scalebar generated from mapeserver, which calculate it with the small extent.
According to my tests, it generate a 0.1% error in a 1/300000 scale. It's not bad. But the best would be not have this error.

2011/2/17 Stephen Woodbridge <woodbri at swoodbridge.com<mailto:woodbri at swoodbridge.com>>
I thought there were some extensive discussions on msCalculateScale and this particular issue in the past. I don't remember if it was on the dev-list or in a ticket. If I recall, it was changed which caused significant regressions and the changed back because of them.

I'm not opposed to a change, per say, but I would like to avoid the issues we had last time. If this is getting into 6.0, this would be a good departure point to make a backwards compatibility breaking change. But once 6.0 is released, we will need to maintain some compatibility within the 6.x family of releases and not introduce major compatibility issues.

-Steve W


On 2/17/2011 10:17 AM, Jan Hartmann wrote:
Interesting. Could this be added to the documentation? It's important
when you are scaling at pixel level. Would it make sense to add an
option for MapServer to treat extents the same way as GDAL?

Jan

On 2011-02-17 16:07, Lime, Steve D (DNR) wrote:

It's due to how MapServer interprets extents. It uses a pixel-center
model so you end up using width or height minus one in computations
because

an extent is from the center of the UL pixel to the center of LR pixel
(you lose the "one" from the outside) . I've attached a graphic I
created some time

ago that illustrates the difference between a MapServer extent and the
model GDAL and WMS use.

Steve

*From:*mapserver-dev-bounces at lists.osgeo.org<mailto:mapserver-dev-bounces at lists.osgeo.org>
<mailto:mapserver-dev-bounces at lists.osgeo.org<mailto:mapserver-dev-bounces at lists.osgeo.org>>
[mailto:mapserver-dev-bounces at lists.osgeo.org<mailto:mapserver-dev-bounces at lists.osgeo.org>] *On Behalf Of *Mohamed
Saâd HESSANE
*Sent:* Thursday, February 17, 2011 4:41 AM
*To:* mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org> <mailto:mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org>>

*Subject:* [mapserver-dev] Implementation of the msCalculateScale

Good morning everybody.
I don't understand how the msCalculateScale function in mapscale.c is
implemented.
As far as I know, the scale is calculated from the relation :
inches_in_screen / inches_in_reallity.
inches_in_screen : depend on the DPI and the width of the map;
inches_in_reality : depend on the map.extent, and the pixel size.
So why the "width-1" at the line 96? For me just "width" is more
correct. Perhaps I'm wrong.
The branches 4.X have just "width" in their implementations (line 125).
So why the center pixel is removed from calculation?


_______________________________________________
mapserver-dev mailing list
mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org>  <mailto:mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org>>

http://lists.osgeo.org/mailman/listinfo/mapserver-dev


_______________________________________________
mapserver-dev mailing list
mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org>
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

_______________________________________________
mapserver-dev mailing list
mapserver-dev at lists.osgeo.org<mailto:mapserver-dev at lists.osgeo.org>
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapserver-dev/attachments/20110217/e587e8a2/attachment.html