Murray,
Did you mean to send to me personally? If not, you can resend to the
list, and I'll resend my reply there.
* Murray Altheim
|
| I'm not sure if that's a good thing or a bad thing, but for my own
| purposes I'm glad LTM remains "XML friendly."
I very much agree. I can't really see any reason to abandon that yet.
| I like the mnemonic of '<' as much as anyone, but OTOH it is
| actually a bit confusing since directedness is not visualized
| consistently by everyone, and this prejudices the visualization
| (i.e., the arrows can point either way in a visualization, depending
| on who you talk to -- it's just an aesthetic thing and there really
| isn't any standard).
I absolutely agree, but on the other hand there is nothing but
aesthetics to guide us here, and so I've use mine to design LTM.
| Well, as I've gotten further along in my own research I've realized
| that a lot of the precision people claim about their projects,
| products and specifications is illusory, even as it is specified,
| much less as it is used.
I wholeheartedly agree.
| Looking back on XTM 1.0 PSI set I thought (out of ignorance) we were
| being pretty canonical, but now I realize we were simply defining
| one set of possible semantics from among the many possible, and
| absent a truly mathematical definition, it was all just prose.
Actually, I've never been able to find anything which defines the XTM
1.0 PSI set in any way at all. The topics are named, and there is some
prose, but the prose does nothing more than just repeat the
names. (The 'topic', 'association', 'occurrence', 'sort', and
'display' PSIs do have more than this, but very little. Sections
2.2.4.3 and 2.2.4.4 just repeat the same prose, as does 2.3.2.) TMDM
expands considerably on this, but uses (wisely or unwisely) new PSIs.
I've always been amazed that nobody has ever commented on this.
| That doesn't buy anything with mathematicians. Our audience wasn't
| mathematicians, though I was happy to see that paper describing XTM
| in formal logical terms. Can't seem to find a reference to that
| now...
I've seen two, but I can't see much reason for rejoicing in either,
really. As Robert Barta said, they are "theories of XML elements",
which is very different from being theories of topic maps.
| I agree though that the difference is big. Earlier this year I'd
| started a discussion with Bernard about creating a PSI set for an
| order relation. I'm currently using one in Ceryle to set the
| direction of ordered relations (as visualized in the graph), but I'm
| in no way satisfied with the prose of my definitions. At some point
| when Bernard and I have some time, and if you or anyone else is
| interested, it'd be good to at least have a public PSI somewhere for
| ordered relations.
I certainly agree. I doubt I'm the right person to publish this, but
I'd be happy to review anyone else's attempts at producing PSIs for
this.
| At some point for all of FOL.
A bit more to chew off, I think :), but it certainly would be good to
have.
| John Sowa has quite a lot of stuff defined in his Knowledge
| Representation book, and Cyc is a good source too. I prefer not to
| have to define anything myself if somebody with more expertise and
| experience has already done so...
Agreed.
--
Lars Marius Garshol, Ontopian <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50 <URL: http://www.garshol.priv.no >
|
