Daniel O'Donnell writes:
 > This last example raises a final question that is perhaps worthwhile
 > keeping in mind--which is 'is your material best suited to storage as a
 > structurally marked-up text file'?

I would suggest that this is not the point. There are three situations
to deal with
 - authoring
 - storage
 - delivery
and XML/TEI might be appropriate for any of all of these, as might a
database. my rule of thumb would be to say that it needs text markup,
not a database, if there is lots of mixed content. looking at Wendy's
sample, I see a fair bit, and I suggest that putting it into a
conventional relation database would be a real pain.

 > If you want your collection to become a real tool for analysis or
 > department-wide standardisation, however, then you might want to
 > consider producing a syllabus database.
I cannot see anything you can do with a database that you cannot
do with a fully-marked-up text. There *is* no difference, its
just a matter of storage and query techniques. Walking through an XML
file with XPATH logic is just as expressive as SQL - the difference
may be in *efficiency*, which should be the concern.

Something like http://exist.sourceforge.net shows how you can store
XML in a relational database, and get the best of both worlds.

 > allow you to do very powerful things--derive a faculty wide
 > bibliography, for example, or compare courses on the bases of their
 > assigned texts or reading lists.
that would be well within the abilities of a corpus of XML files, and
a bit of XSLT or the like

 > problem over for you.  Syllabi are very much like customer invoices,
 > which are the example most first year relational database textbooks seem
 > to choose for their first complex normalisation exercise.

invoices tend not to say ""Mosley, the whole darn book" :-}

Sebastiabn