To concretize this discussion a bit by focusing on what I know -

My archive, Washington University, is certainly *not* ready for a world
that would make Minsky or Gelerntner happy. Our data, right now, is not
ready for anything but the most basic machine-to-machine communication,
and our interfaces are far from allowing transparent access to our data.
But we're moving, slowly, in that direction, and there is a good
explanation why this is the case now.

Since the hardest work here at WU involves legacy descriptions, and that's
also the bulk of the work, the "loose" nature of the DTD has tremendous
value to me. Usually I have a sets of descriptive documents done by a
variety of practicioners, whose preparation of the documents (meant to be
read from paper) may have varied widely. Sometimes, even within a time
frame, descriptive practices for the same type of document vary. Anyhow,
because of this, it's really helpful to have a variety of ways of doing
things available. Each of these groups of finding-aids have their own
internal logic, and (with our guidance) our vendor tends to make the same
choices about the different groups finding-aids, but not all the documents
as a whole. But as long as I have groups with their own internal logic, I
can conceivably write local transformations to normalize and standardize
the data that our vendor prepares for us. I see no way around this step.
As Michael points out, this is going to be a maintenance headache, but
until the dust settles there seems to be no other way to move forward,
and, anyhow, commitment to upkeep goes along with all sorts of electronic

I know from my own practice here at WU that not every question regarding
how to create an EAD document has a single answer ready at hand (though
the application guidelines and the cookbook are very good). More and
better answers are emerging as we continue to re-examine practices with
these documents. But certainly the choices made in the legacy documents
could not have foreseen emerging information standards. Even for
contemporary documents, the information standards that we use evolve.
Anyhow, there are still a huge number of legacy documents out there, and
even the best education on standards to enhance ongoing descriptive
practices and tagging practices won't change that.

I find that my conscience and I get along better if I carefully document
the choices I make with these legacy documents regarding guideline sets,
procedures, document structures and etc. and make them consistently. These
choices usually end up being a compromise between the possible and the
ideal based on the fact that I have limited resources and time.

But if I'm at least consistent in my choices, what we or a vendor under
our direction does can be undone and remade if it becomes necessary.
Regular expressions's, XSLT, good documentation, and some work should make
it possible to transform and normalize consistently done local documents
as the dust settles and standards firm up. EAD, as it is now, enables this
approach to work.

Speaking of regular expressions, at times I heavily use regular
expressions directly on text-files for normalization and data-massage.
Perhaps if the SAA education office is looking for a sense of where to go
with this kind of thing there might, perhaps, be some discussion of why
and how to use regular expressions along with a curriculum on XSLT?

But even now our EAD documents can easily interact on a limited level with
the much more standardized MARC world. For instance, the cookbook
stylesheets have prepared HTML finding-aids that can be indexed, creating
MARC records through using OCLC's CORC system. Or, once the document is
edited for content, one could imagine writing a transformation for bumping
down the XML tree and producing XML MARC records for series, subseries,
even items (when they are there) using LOC's DTD.

Michael's e-mail underscores for me that many archives are very much where
ours is - struggling to work with emerging standards, trying to maintain
consistency and control of information, and working as best we can in the
real world while moving towards something better.

Chatham Ewing
Washington University