[sc34wg3] Conformance

[Kal Ahmed]

Lars Marius Garshol wrote:

>* Kal Ahmed
>|
>| I agree with Lars Marius. [...]
>
>Thanks for saying this explicitly. 
> 
>| Round-tripping the XML interchange syntax of the topic map is not
>| really a test of conformance. 
>
>Well, it's a requirement for it
>
That depends on what you mean by round-trip. IMO a round-trip should 
preserve all aspects of the input, including ordering, attribute values 
that are not part of the SAM model and so on. I presume that this is not 
your definition...

>, and I hesitate to require support for
>CXTM in order to conform. In fact, if you can roundtrip through
>someone's tool you can use an XTM2CXTM converter that is known to be
>conforming to test that tool, even if it does not support CXTM.
>

Thats a good point. Presumably the XTM2CXTM converter would simply 
perform the necessary ordering and serialisation in canonical XML ? Such 
a converter would certainly lower the bar for apps seeking to claim 
conformance, which is a Good Thing. Now, who volunteers to write it ? ;-)

Cheers,

Kal