Dear Marjorie,
I would be interested in a feedback of your practice: do you, when you teach the TEI, feel the need / manage to integrate notions about other tools and technologies of the XML family? And if so, how do you do that (proper introduction to XSLT? use of packaged tools like the v-machine? etc.).
when I teach the use of TEI in course of mansucript description and/or digital editions a very important notion is the one that encoding has to be done with not only looking at what is there (e.g. in the source) but as well what the result of this exercise would have to look like.
"Don't encode what you don't use for research or presentation!"
Only if one knew what s/he wants to achieve "proper" encoding is possible. In order to know (or at least guess) what you *can* achieve I think its necessary to know a bit about the processing (e.g. XSLT) or the publishing process (e.g. DB plus XQuery).
But: An as strong point as the previous is to take into consideration that it might be helpful to (re!)establish the division of work: One to encode data and do or at least prepare research on the data and let someone else take responsibility for publishing both original data and research results. Although it has become common practice that researchers have to publish their work on their own ("Word is doing such good work"), in the field of DBs and X-technologies the gap between good contents and good "publication" is even deeper than it was in the age of print. (We only had forgotten that good print needs something more than just Word and were left with this task!) Thus the presentation to XSLT or DBs might be rather limited, more like just a glimpse to know what's out there.
Courses which show these aspects are: (information mainly in German, sorry)
http://www.hab.de/bibliothek/sammlungen/hzdfg/SCRIPTO/index.htm http://www.i-d-e.de/spring-school-2011
I consider all this is valid for all XML-based work not only TEI.
Best, Torsten