Hilde Bøe wrote
> In the establishing of Henrik Ibsen's Writings' main texts,
> the project is examining all the text witnesses from Ibsen's
> lifetime. We encode the bibliographic information on all the
> text witnesses to a work in the sourceDesc> element of
> the main text's <teiHeader>.
> We have chosen to use <sourceDesc> instead of <witList> because
> we prefer to record all of the bibliographic information only once (and
> in the same place) and because we prefer to keep the witness list
> information in the TEI Header since we consider it to be meta data.
> Still, we would like to know how others have solved this. How many
> are using the <witList> element and where do you prefer to put it?
> Are there other solutions in use? Does anybody have objections to
> our solution? We also wonder why the <witList> element is not
> included in the <teiHeader>, which to us seems to be the obvious
> place for this kind of information.
This all strikes me as plausible and appropriate to the situation described,
but to get a better angle on the broader question, consider the different
case where an existing printed critical edition is being digitally encoded.
I can take one such down from my shelf right now.
Immediately after the "Editor's Preface" it has a "List of Manuscripts
consulted with Key to Sigla". And what follows almost marks itself up as a
<witlist>. But once so marked up it clearly belongs in the <front>, to be
rendered after the Preface. It isn't a description OF the source(s) being
encoded, it's a description WITHIN the source being encoded (albeit
describing the original editor's sources), hence in this case not metadata,
and it would be misplaced in the teiHeader.
The GLs also seem to envisage situations where the <witlist> (or fragments
thereof) does not necessarily correspond so straightforwardly to an
identifiable discrete part of the printed text and may not be destined to be
rendered in its immediate form at all. In such instances its function is to
"simplif[y] the automatic processing of the apparatus" (P4 List of
Elements ). Though the example given (P4 Ch 19.1.1) seems to embody a
somewhat desperate view of "simplicity" (using the witlist to deduce from a
list of variant readings in a given <app> the identity of those sources
which agree with the base text) one gets the point. So is this now metadata?
I'd say not. It is a kind of convenient gathering together of things that
are already present in the text itself. If I were not determined to avoid
trouble so close to the weekend, I would venture to term it "paradata"; but
I am, so I won't, and please pretend I didn't.
With that perspective established, let's look back to the situation from
which the questions were posed. Not the digitisation of an existing printed
book, but the origination of a new edition in digital form, with a variety
of printed editions as its witnesses. Now in this case there are indeed
clear overlaps between what a metatextual Source Description is meant to
cover and the function of a list of witnesses. So quite a lot to pile in the
metatext pan of the balance. But the other pan won't remain empty and aloft
for long. After all, even if this is to be a purely digital edition, readers
will expect to see a list of sources, probably with bibliographical
commentary and a key to abbreviations used when indicating variants etc. And
that is likely to be in the digital counterpart of the front or back matter
of this edition: so maybe not such an obvious candidate for the teiHeader
after all. So the arm of the balance may well now be pretty horizontal,
waiting to be tipped by two-cents worth of good old encoder's hunch laid one
side or the other: better than letting a blast of hot air from here in the
UK do the job, anyway.
Bon weekend à toutes et à tous!