Re: [Ecrit] Profiles (was: Consensus Call) - namespaces

[Henning Schulzrinne <hgs at cs.columbia.edu>]

On May 9, 2007, at 9:33 AM, Andrew Newton wrote:

> On May 7, 2007, at 11:08 PM, Henning Schulzrinne wrote:
> > I don't see what hints you can offer in this particular case. As  
> > far as I know, we have not attributed any semantics to the  
> > ordering of the elements. In general, I think it would be a bad  
> > idea to do so, given that the schema considers both orderings  
> > equivalent.
>
> Consider location profiles that are not yet defined.  Which one is  
> this?

Obviously, there's no way to tell. But I thought we earlier agreed  
that unknown GML would get no profile label. The important part is  
that the receiver can recognize the GML. If it doesn't recognize the  
combination of point and circle, it is going to have to throw in the  
towel, with or without a profile label.

> <gml:Point srsName="urn:ogc:def:crs:EPSG::4979">
>   <gml:pos>-34.407 150.883</gml:pos>
> </gml:Point>
> <gs:Circle srsName="urn:ogc:def:crs:EPSG::4326">
>
> </gs:Circle>
>
> Of course, if you are suggesting that future profiles the glue  
> multiple elements together such as this would need to define their  
> own schema to properly define them, then I consider that a fair  
> design choice.  That does mean that the current method of 3825  
> compound location defined in pdif-lo-profile would need a schema  
> that wraps the gml and civicAddress.
>
> Also, why would the LoST server need to identify profiles then?  It  
> should use the same XML namespace mechanism.

No, it should recognize it by the the combinations of elements that  
it supports. If it sees something it can't recognize, it has to  
return an error.

The only question is whether it should ignore elements that it  
doesn't understand, as in

<Point>...</Point>
<Blob>..</Blob>

Should it ignore the blob and just go with the point?

The compound locations in draft-ietf-geopriv-pdif-lo-profile-07 have  
well-defined rules as to what should come first and what can be  
included, so I fail to see the problem. (Again, we do need to  
coordinate that whatever is legal in LIS, according to the phone-bcp  
or whatever other document, is also legal in LoST, to avoid problems.  
That, however, isn't fixed or affected by a profile indication.)



> > Or are you suggesting that the receiver treat this only as a hint  
> > and ignore the label if it contradicts the XML?
>
> Yes.

As long as we specify this as a 'MAY insert as a hint' and 'receiver  
MUST ignore if it contradicts the XML', I'm not too hung up about  
including it. The use seems limited, but harmless.



> -andy


_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit