Information Research Weblog

I agree with Tom. I have argued much the same elsewhere. The problem(s)
with any free text system -- whether one is asking authors/editors to
provide keywords or to undertake an extensive xml markup is the very
free nature of the process. The process lacks any modicum of the use of
controlled vocabulary. In order to generate anything approaching
relevant retrieval, one would need to have access to multilingual
thesauri well in excess of what the EU, for example, has produced.

Second, any cataloging/indexing scheme left to author discretion runs
the risk of assumptions, time constraints, etc. In order to capture
full semantic meaning using automated mark-up systems, I suspect all
terms would have to indexed together with the aforementioned thesauri.

Thus as Tom suggests not only are we in need of technical models, we
need also consider economic models.

Finally, I question the very exercise. Given the highly fluid nature of
the WWW, what are the implications for any mark up system when the
underlying native document may be changed virtually at the will of its
creator?

Wallace Koehler