[Mapserver-users] WMS BBOX vs. Actual Bounds

[Arnulf Christl]

Paul Spencer wrote:

> For the record, I believe that the ability to return a map with 
> non-square pixels is one of the few areas (or the only one?) in which 
> MapServer DOES NOT comply with the WMS specification and, in fact, 
> prevents anyone from seeking WMS compliance certification for MapServer. 

Hi,
we run accross this problem all the time when integrating MS with other
WMS - and it always takes a lot of time to explain the implications. The
WMS spec is very specific about this (as Ian stated) and it does make
sense when you think about it long enough.
Most proprietary mapserver in the beginning had the same problem and we
came accross at least 5 different ways to "cheat" the map extent (lower
left, upper left, center center, center top, center bottom).
We hope that it will be implemented correctly (as specified in the WMS
spec) in UMN MapServer soon because many of the proprietary packages out
there already have done it and it gets more and more complicated to
integrate these compliant WMS with the UMN MS.
Probalby it will cause an uproar of all those users who currently use
the "wrong" implementation and got used to it or have adapted their
software to manage this problem (as we have done).

> Note I am not expressing an opinion either way here ;)

We do express an opinion: Better change it now as someday (soon) it will
have to be done anyway (& it doesn't matter whether we think its nicer
or not).

Cheers, Arnulf.

>
> Cheers,
>
> Paul
>
> Ian Turton wrote:
>
>> At 08:56 08/01/2004 -0500, Ed McNierney wrote:
>>
>>> Ian -
>>>
>>> I can't agree that this is poor behavior.  MapServer should also be
>>> expected to return a map in the PROJECTION requested, which is
>>> impossible under your scheme.  In WMS mode the SRS is part of the
>>> request and should be respected just as much as the WIDTH, HEIGHT, and
>>> BBOX parameters, shouldn't it?
>>
>>
>>
>> Indeed - they are all important. However the WMS spec is very 
>> specific about how to handle this case and it says you must honour 
>> all the parameters. Otherwise people like the origianl poster are 
>> left not knowing what they got back.
>>
>>> If you want the image "distorted", you're not looking for a map - 
>>> you're
>>> making a picture.  If you can define the distortion in a
>>> cartographically accurate way (e.g. a picture), then it is indeed a 
>>> map;
>>> but you can also provide MapServer with the definition of that
>>> projection and cause it to "distort" the image the way you want.
>>
>>
>>
>> For example, suppose I'm displaying my picture/map on a display with 
>> non square pixels?
>>
>> Ian
>>
>> Ian Turton, Director, Centre for Computational Geography, University 
>> of Leeds, Leeds, LS2 9JT
>> http://www.geog.leeds.ac.uk/people/i.turton 
>> http://www.ccg.leeds.ac.uk http://www.geotools.org
>> +44 (0) 113 343 3392 fax: +44 (0) 113 343 3308
>>
>> _______________________________________________
>> Mapserver-users mailing list
>> Mapserver-users at lists.gis.umn.edu
>> http://lists.gis.umn.edu/mailman/listinfo/mapserver-users
>>
>