Your thoughtful reply confirms my sense that the answer to your original
question has relatively little to do with technical questions about TEI
and a lot more with how you think about appropriate ways of recording and
recognizing editorial labor.  With some ingenuity people can (and will)
put almost anything in almost any element, and over time the appropriate
content of teiHeader or sourceDesc is shaped by what people put into it.

I talk a fair amount with librarians, and when they talk about the
teiHeader, it's my sense that they tend to think of it as a kind of MARC
record--but not as good.  This may be quite misguided, but it's probably
a fact of usage, and in the long run Usage is King.

There is a way in which it doesn't matter whether all this stuff goes
somewhere in the header or some other place: after all, it's part of an
electronic file and with the appropriate programming instructions can be
summoned up and displayed in many ways.

On the other hand, there may be a way in which the major categories of
the TEI dtd set up different kinds of "documentary space," each with its
own symbolic resonances. And my hunch is that you may be too modest about
the placement of scholarly information that deserves, so to speak, a
corner office rather than a back office.



On Mon, 4 Aug 2003 14:34:14 +0200 Hilde Bøe wrote:

> Sorry for the late reply - I've been out of office a few days. I
> would like
> to comment a little on the two replies to my email and to clarify a bit
> about Henrik Ibsen's Writings.
>
> The forthcoming Ibsen edition is not a solely electronic edition, it is to
> be published both in print and electronically. We are now preparing the
> printed edition (the first volumes will be published in 2005). The
> electronic edition will not be published until the whole project finishes
> in 2008. We have just begun to prepare our files for print publishing, and
> it was in this process our <sourceDesc>/<witList> problem appeared.
>
> Until now we have been encoding diplomatic transcriptions (with no
> editors'
> interventions except for the encoding itself) of all the text witnesses
> from Ibsen's lifetime, including both printed and handwritten source
> material. We have documented our work extensively in the <teiHeader> and
> have taken care to keep the <text> element reserved for the source text
> only with so few interventions/additions as possible. Because of this, we
> probably haven't got used to the thought of putting "our own stuff" into
> the <text> element, but are still thinking of ways to incorporate this in
> the <teiHeader>.
>
> We are still in the middle of the process of deciding how to best organise
> our material for the two forthcoming editions. The printed edition will
> consist of separate volumes for Ibsen's text and the introductions and
> commentaries. The electronic edition will include the material from the
> printed edition as well as the full text of all the examined text
> witnesses. Here the commentaries will be linked to the texts, and
> searching
> and collating text witnesses will be possible.
>
> The <text> elements of the files containing Ibsen's text will therefore
> probably not contain anything else from the editors than the critical
> apparatus. Introductions (with discussions on text witnesses and
> manuscript
> descriptions among other things) and commentaries will probably be encoded
> in separate files. We believe that this is convenient since this material
> will be printed in separate volumes. We need to decide how to incorporate
> several works in a single file, but this has yet to be discussed
> within the
> project.
>
> At the current stage of our work (with all text witnesses in separate
> files) the possibility to check and validate the sigla used in the
> apparatus just by making a minor modification in the DTD and exploiting
> excisting information in <sourceDesc>, makes this the best solution
> for us,
> but the replies from Michael Beddow and Martin Mueller have convinced
> us to
> reconsider using <witList> instead of our own <sourceDesc> solution
> in the end.
>
> This said, we still feel very strongly that the <teiHeader> is much more
> than a catalog card, and that it should be used to document as much as
> possible of the work done to the file, whatever aspects it may concern.
>
> Best regards, Hilde Bøe.
>