[postgis-devel] WKT and Wikipedia

[Paul Ramsey]

Be liberal in what you accept and strict in what you emit.  The
question is, whose rules should we adhere to?  I guess, really the new
rules are the only super-official ones, since the old SFSQL never
actually went beyond 2D.

P.

On Mon, Jun 30, 2008 at 3:25 PM, Kevin Neufeld <kneufeld at refractions.net> wrote:
> That is a good question.  Since it looks like we going to break people code
> anyway ... to be true to the spec, I assume this also means that the parser
> should error out if someone tried to supply 3 or 4 dimensions without using
> the POINT Z, POINT M, or POINT ZM syntax.
>
> We could try to be backwards compatible and continue to allow the use of
> 'POINT' for 3D/4D coordinates, but as Martin pointed out, we would no longer
> be compliant with the standard since the OGC standard itself is not
> backwards compatible.
>
> Choose your poison, eh?
>
> -- Kevin
>
> Paul Ramsey wrote:
>>
>> Yes, handling POINT M, POINT Z, and POINT ZM is a good thing for 1.4.
>> That's an addition to the parser (I can put my new lex/yacc book to
>> use :). The hard question is what to *emit* in response to an
>> ST_AsText() :)
>>
>> P.
>>
>> PS - those are the ISO wkt forms too.
>>
>> On Fri, Jun 27, 2008 at 4:42 PM, Kevin Neufeld <kneufeld at refractions.net>
>> wrote:
>>>
>>> So, the 1.2 OGC specs require that the WKT of Z, M, and ZM be:
>>> POINT <point text>
>>> POINT Z <point z text>
>>> POINT M <point m text>
>>> POINT ZM <point zm text>
>>>
>>> PostGIS has:
>>> POINT <point text> | <point z text>
>>> POINTM <point m text> | <point zm text>
>>>
>>> To me, this means PostGIS is not entirely OGC compliant. (ie. PostGIS
>>> can't
>>> parse an OGC compliant WKT string)
>>>
>>> postgis=# SELECT 'POINT ZM(0 0 0 0)'::geometry;
>>> ERROR:  parse error - invalid geometry
>>>
>>> Same goes for the textual representation of other geometries.
>>> ie. LINESTRING ZM (...)
>>>
>>> Should we add this to the TODO list for 1.4?  Or is there a specs gray
>>> area
>>> ... that we're compliant enough.
>>>
>>> -- Kevin
>>>
>>> Martin Davis wrote:
>>>>
>>>> Actually the latest SFS 1.2 spec does include a spec for 2D measured, 3D
>>>> &
>>>> 3D measured WKT.  See here:
>>>>
>>>> http://www.opengeospatial.org/standards/sfa
>>>>
>>>> (This is also referenced on the wiki page)
>>>>
>>>>
>>>>
>>>> Kevin Neufeld wrote:
>>>>>
>>>>> I just came across this page:
>>>>> http://en.wikipedia.org/wiki/Well-known_text
>>>>>
>>>>> I think this page is incorrect, misleading, and contains typos.  It is
>>>>> my
>>>>> understanding that the OGC specs on WKT does *not* include 3D and 4D
>>>>> definitions.  It is a PostGIS extension to use Z, M, or ZM. Or am I off
>>>>> on
>>>>> this?
>>>>> (http://portal.opengeospatial.org/files/?artifact_id=829) (section
>>>>> 3.2.5.2)
>>>>>
>>>>> I don't have an account to correct this.  Does any here have one?  Do
>>>>> we
>>>>> care?
>>>>>
>>>>> Should the new PostGIS docs reference such pages, or point to only the
>>>>> OGC specs?
>>>>>
>>>>> -- Kevin
>>>>> _______________________________________________
>>>>> postgis-devel mailing list
>>>>> postgis-devel at postgis.refractions.net
>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>>
>>> _______________________________________________
>>> postgis-devel mailing list
>>> postgis-devel at postgis.refractions.net
>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>
>> _______________________________________________
>> postgis-devel mailing list
>> postgis-devel at postgis.refractions.net
>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>