I think one way of answering Trevor's call to focus on translation is
to respond with more focused community tool building. The desire of
the TEI to be software agnostic is laudable, but taken too far it
simply isolates the standard. We have too few widely-used tools for
dealing with markup once it's produced. Most HTML encodings aren't
tested with W3C validators, but by sticking in two or three browsers
and seeing what comes up. This is, of course, the root of some
problems in HTML, but it's also evidence that what most want/need is
not markup that is theoretically valid, but markup that _works_.
What if the TEI now devoted its energies and financial resources to
sponsoring community open source tool building efforts. As the
community begins designing a few "browsers" (not only ones for simple
presentation but also for search and visualization), I think we will
better understand how to restrict (and possibly expand) our standard.
There have been several TEI tools developed by small, relatively
closed development efforts in the past (I've been a part of a few),
but I'm unaware of any multi-institutional open source effort not
associated with (and limited to the timing of) a particular grant.
In short, if the community looked more like the Mozilla foundation
than the W3C (maybe only for a bit), I think we'd better understand
our own needs.