[georss] GML georss namespace

[Knoth, Brian D.]

The problem occurs when a client developer then wants to use the normal
process of compiling a Java binding from the schema to deserialize the
georss data, and it won't work. They will need to modify the schema and
hope that no georss producer is compliant with it. 

I'd like to see the application schema that is posted changed to
reflect what the georss namespace is really expected to be. Oh, and by
the way, while whomever is doing that, could they also add optional
begin/end time attributes to the georss:where element?

Also, my experience with the Microsoft Virtual Earth georss
functionality is that it does appear to validate against some georss
schema. It is very fussy.

Brian

-----Original Message-----
From: Gregor J. Rothfuss [mailto:gregor at apache.org] 
Sent: Tuesday, March 20, 2007 7:49 AM
To: Knoth, Brian D.; georss at lists.eogeo.org
Subject: Re: [georss] GML georss namespace

Knoth, Brian D. wrote:
> According to the current GML georss application profile here:
> [http://www.georss.org/xml/1.0/georss.xsd]the target namespace for
the
> georss schema is "http://www.georss.org/georss/10", however, every
> example I have seen and code we have written uses the namespace
> "http://www.georss.org/georss". These instances do not validate
against
> the application schema, so I'm wondering what gives? 

good catch. i have never seen /10 in the wild, so it is safe to say
that 
the schema is wrong. while it would have been a good idea to version
the 
xmlns, that ship has sailed.

> Which is the real namespace that should be used and are most clients
of
> georss validating against the application schema?

i very much doubt it. one of the highest quality parsers on any
language 
is called ultra liberal feed parser in recognition of the sorry state
of 
most rss feeds out there.

http://feedparser.org/

in light of that, i don't think xsd validation is a viable strategy,
sadly.

-gregor