This is an interesting discussion, and there is no right or obvious
answer. Perhaps it's best to think about it as the problem of metadata in
digital texts. In a print world metadata are largely bibliographical. The
teiHeader nods in that direction and acts as a kind of catalog record for
the text that follows. The XML structure of the text is a kind of metadata
that is stored "inline." I take it that one could model it as a kind of
stand-off markup. Linguistic annotation is a form of metadata that can be
stored inline or in a stand-off mode.
In one literary genre, drama, a prosopography in the form of 'dramatis
personae' has for centuries been part of the genre and is appropriately
encoded it as a part of the text, although one could have a discussion
whether it should be part of <front> or <body>. I think I have seen both.
Linguistic or prosopgraphical metadata may be part of a single text, but
it is just as likely that they are a form of authority list to which many
texts refer. A lexicon of all words in a corpus is, if you will, a
'prosopography' of all the words in that text, and it will include data
about the word that are not found in any particular text but may emerge
from their aggregate. Alternatively you can think of a prosopography as a
lexicon of people.
What to do with such corpus-wide metadata (and how to make them accessible
to users in easy and powerful ways) is likely to be an increasingly
important problem in the management of digital texts.
On 8/16/11 10:46 AM, "James Cummings" <[log in to unmask]> wrote:
>Like Torsten I've always felt nervous about putting these in
><back> because they are not actual appendices to the text I'm
>transcribing (ok, I recognise this distinction may not exist for
>born digital documents, where they might be considered
>appendices). So I instead have convinced myself that they are
>supplementary information I use to describe the source, and thus
>put them in sourceDesc where they are conveniently allowed. I'm
>not arguing that is correct, just describing what I've done in
>the past. I have also, of course, just stored them separately
>and pointed *to* them as separate files, not feeling any need to
>suck them all in (virtually or otherwise) into a single file.
>My two pence,
>On 16/08/11 16:13, Martin Holmes wrote:
>> HI Gabby,
>> Typically I keep these in external files. When querying an XML database
>> to get back a single TEI file, though, I'll often import all the
>> <person> or<place> elements which are linked from that particular
>> and place them in<back>, so I have a single internally-coherent XML
>> structure for processing with XSLT. I have also put them in the
>> <teiHeader>, though.
>> On 11-08-16 06:42 AM, Gabriel Bodard wrote:
>>> A question for the community about usage:
>>> Where do you (in your own local usage) place, within a TEI document, a
>>> prosopography (listPerson) that is neither a castList nor a
>>> description (particDesc), nor a list of scribes in the manuscript etc.;
>>> or a gazetteer (listPlace) that is not a setting description for your
>>> document; or a listNym, listOrg, listEvent? In other words, if the
>>> is serving as a vocabulary or external table to your digital corpus,
>>> rather than a manuscript description or catalogue metadata.
>>> My instinct is that such a list doesn't belong in the teiHeader, but
>>> perhaps in the tei:body or tei:back of a text, depending on the
>>> structure of the edition being created.
>>> What have others chosen to do?
>Dr James Cummings, InfoDev,
>Computing Services, University of Oxford