[topicmapmail] PSIs?

[Lars Marius Garshol]

* Murray Altheim
| 
| Okay, I can certainly state that the "PSI" URLs are the ones to use,
| since they point to the XTM document.

That would help. It would mean that you don't conform to the OASIS
recommendations, though.

| If you've included that topic map (datatypes.xtm) in your system, it
| shouldn't matter, eh? 

I don't think it's appropriate to assume that people must do that.

| The PSI in datatypes.xtm has subject identity with the XSD sID. 

Unless it also has the PSI URLs in there you'll have to load the file
over HTTP to get the equivalence declarations between the PSI and
"sID" URLs. That won't work in all cases, of course.

* Lars Marius Garshol
|
| Why not make the <subjectIndicatorRef/> in the XTM document point to
| the URI derived from the XSD Recommendation? Then you have no loop,
| and everything should be fine. You can have a single URI and simplify
| your document quite a bit.
 
* Murray Altheim
|
| Well, that's how it actually is right now. 

In that case, why not just ignore the "PSI" URLs and only use the
"sID" ones? Then you only have a single set of URIs and all is well.

| I wouldn't want to do that, since the XSD spec is *semantically*
| the canonical definition for what these PSIs mean.

That makes sense to me.
 
| I think all this is resolved by (as you say) simply stating that
| within topic map environments, the "PSIs" as listed are the ones to
| use. The "sID" are noted as being subject identifiers within the
| topic map and as links to definitions, but shouldn't be used (but
| could be if the datatypes.xtm is brought into a system of maps).

Well, the whole point of creating published subjects is to define a
*single* URI that identifies each subject. This means defining two for
every single one. Surely you agree that that's not good?
 
-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >