stability of TEI guidelines

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

stability of TEI guidelines

Michael Sperberg-McQueen
Richard Goerwitz asks whether the TEI guidelines are stable enough to
bother writing software for, yet.  To this, of course, there is no
single answer:  it depends what you are trying to accomplish, how much
you enjoy writing software, and how you feel about throwing away your
rough drafts.  My personal take on this issue is that the guidelines are
stable enough to begin working with software, and unstable enough that
if I were developing software I would expect to make some changes to the
software after each revision of the draft, and I would never tell anyone
it was stable until after 1992.

To be blunt:  Anyone who needs something which won't change soon
shouldn't rely on the TEI's current draft.  (That's what it says in
section 1.4.)

But for many purposes, this is a fine time to start.  The major problems
facing any stand-alone processor for TEI files are probably SGML-related
in any case (recognizing tags, handling entity references, etc.) and
won't change unless SGML changes.  Anything you write to handle specific
tags is subject to more change, and if I wrote a bunch of stuff to guide
processing by an SGML parser (for example), I'd expect to throw a lot of
it out eventually -- half because the tags change and half because I
will change my mind about how to process things.  I wouldn't let that
prospect deter me from beginning to work with the current draft.  In
fact, were I a software developer, I would want to begin work with the
draft as soon as possible, so that I could argue for changes to make the
TEI format more tractable for the things I want to do.  And of course,
the future drafts are likely to be similar enough in spirit that I would
expect to be able to transfer most of the understanding gained from
processing files encoded according to the earlier drafts.

So no, I don't think Bob Kraft's student is premature in his plans.
Most of what he does should be usable with any future draft.  of the TEI
guidelines.

-Michael Sperberg-McQueen
 ACH / ACL / ALLC Text Encoding Initiative
 University of Illinois at Chicago