[topicmapmail] stupid question?

[Lars Marius Garshol]

* Lars Marius Garshol
|
| I know I've said this before, but people *really* need to stop
| thinking XSLT is a real solution for topic map publishing.

* alexander johannesen
| 
| Hmm, isn't that a bit broad and sweeping? Or do you mean XSLT 1.0
| specifically?

No, I mean precisely what I say.
 
| I'm currently in the works of writing an XSLT-extension for
| TM-handling, because I do a lot of work in XSLT already. 

Wrong reason, I think. :)

If you had no budget and didn't want to look at the free topic map
engines, or found them unsuitable for some reason, that would be a
good reason.

| XSLT, dealing with XML is - as you know - a very slick solution to
| certain problems (albeit not all), 

It is. I like XSLT a lot, but XSLT doesn't work very well for
structures that are not XML-like, and topic maps are not at all
XML-like. They are graphs and consist entirely of references and
collections, the two things that XSLT handles least well.

Also, topic maps are entirely based around identity, which is a
concept that is completely lacking in XML, and in fact quite difficult
to handle. (Ask any XML database vendor.)

| [...] so where the orogonal spec lacks features (like simplified
| association handling) they can be added through this extension. Is
| this what you mean by "library"? 

Yes. But I doubt a library will help much. Jonathan Robie tried to
show how XQuery function libraries might help, but I wasn't
convinced. 

| (And using Nikitas solution from the "Topic Map" book is not an
| argument; there are several things there that have far more elegant
| solutions, and I guess his code wasn't meant for more a
| proof-of-concept that "do it this way")

I haven't seen that chapter, so I can't comment on it.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >