However, we nearly always require, in an IETF document sense, that extensions be defined in document and/or a registry, and that they have at least a modicum of IETF review before adding them to a registry. We do this to foster interoperability and to prevent dumb things from happening.
A lot of protocol mechanisms work the same way - for example, we often have ABNF that is more permissive than the document actually defines.
In this case, some of use are saying that even though anyone can already add anything they want to a PIDF, and get a validated XML document, don't do that.
Brian
On Sep 8, 2010, at 7:52 PM, Thomson, Martin wrote:
> James Polk writes:
>> Brian and I (at least) are stating that registries prevent collisions
>> and increase interoperability by understanding what a given element
>> means. If some random other SDO or group of guys in a garage decide
>> to define a <cl:HMO> element that isn't in line with RFC 4119, 4776,
>> 5139 *and* IANA -- they are increasing the likelihood of
>> miscommunications.
>
> James,
>
> Namespaces in XML have inherent protection against collision. The only viable concern is the interoperability one. If you want that; we have a registry.
>
> --Martin
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
