The EPSG Contract

[Paul Ramsey]

Well, the news from OGC on the WFS/GML issue seems to be to accept  
the implemented facts on the ground for older representations of epsg  
codes ("EPSG:XXXX") and enforce the "consistent" behavior for newer  
representations of epsg codes (urn notation).  And the WMS 1.3  
specification remains as writ, so when Mapserver implements 1.3 it  
will need to follow the specification, such as it is.

This implies, incidentally, some extra infrastructure at the backend,  
because the current Mapserver source of EPSG CRS information, the  
proj4 library, does not include ordinate order in its backend epsg  
database (the epsg file).  That leaves a pair of unpleasant  
implementation paths: add a whole new concept into proj4, and then  
have mapserver require the new library version in order to support  
WMS 1.3; or do a hack solution and simply assume all +proj=longlat  
combinations have reversed ordinates, unfortunately the hack will  
omit other cases, like several south african and some european  
projects, that also have a northing/easting order.

So long story long: no news :)

P

On Jan 7, 2006, at 8:30 AM, Arnulf Christl wrote:

> Any news here? Sorry if i missed it. We want to support 1.3 and are  
> trying to get an overview of which services will support it in  
> which way. We are the poor last layer and have to deal with  
> whatever you spit at us.
>
> Best, Arnulf
>
>
> Paul Ramsey wrote:
>> I have also tried the Intergraph WFS and it is in easting/ 
>> northing, so
>> that makes 3 so far:
>>   Mapserver
>>   Cubewerx
>>   Intergraph
>> Will be interesting to hear what the wfs list says.
>> Frank Warmerdam wrote:
>>> On 5/19/05, Paul Ramsey <pramsey at refractions.net> wrote:
>>>
>>>> Is it so bad to suck it up and follow the standards consensus,  
>>>> no matter
>>>> how silly? Sometimes bad decisions are made, but walking away  
>>>> from the
>>>> process destroys all the good decisions too.
>>>
>>>
>>>
>>> Paul,
>>>
>>> I do see your point, and I have certainly accepted that when  
>>> MapServer
>>> implements WMS 1.3 it will have to address this issue properly.   
>>> However,
>>> the WFS stuff has existed for quite a while (as has GML) but this  
>>> issue
>>> is only being raised now in the WFS/GML context.  So we are left
>>> having to back-fix without any obvious indicate of what is really  
>>> the right
>>> thing to do.
>>>
>>> I *suspect* that at least for GML 2 the coordinate system does not
>>> control order of values in coordinate tuples.  I'm sure I have heard
>>> previous detailed discussios to this effect.  I am not at all  
>>> sure of
>>> the situation with GML3.
>>>
>>> Hopefully we can isolate these issues to GML3 and WMS 1.3+ so that
>>> a simple version check will suffice to make the decision clear.
>>>
>>> Best regards,