Mon, 28 Apr 2003 09:20:44 +0100
Lars Marius Garshol wrote:
>* Robert Barta
>| I think it is quite simple:
>| 1) you create some implementation of what you think is a TM
>| implementation, in that you create your databases, internals,
>| 2) to "prove" that your application is "SAM conformant" you simply
>| define a mapping of concepts of SAM (and operations if we had
>| those) onto your API (or your API concepts).
>Correct me if I'm wrong, but I think Jan already accepted this, and
>what he is asking is whether step 2) has any value here. Personally, I
>don't think it does, and I think the conformance requirements of the
>syntaxes, APIs, and query/constraint languages will be all we will
>need, but the opinions of others would be welcome on this.
I agree with Lars Marius. Unless/until there is a common API, I think
SAM conformance in terms of an API are pretty meaningless. What is not
meaningless though are the operations that the SAM requires a topic map
processor to perform and validation that a processor does indeed perform
those operations is probably best done not by inspection but by testing
the application against a conformance test suite. So SAM + CXTM +
conformance test suite is needed to prove the level of conformance which
I as a user would expect from an application, and to which I as a
developer would build my topic map processing software.
>| [ Hope I got it right in this abridged form ]
>| True, but it will be more difficult to fool the outside if I ask
>| your application to round-trip an XTM document.
>Exactly. So why have your step 2) above?
Round-tripping the XML interchange syntax of the topic map is not really
a test of conformance. XTM->processing->CXTM should be the way to
validate a topic map processor.