Many thanks to all who replied on and off-list with this. I think 
Torsten's right that there are different ways to handle this, and that 
depends a bit on the status of the identifiers / accession numbers 
you're recording. (In my case, the multiple fragments have now been 
brought together and may even be recorded together in a database, but 
the original accession numbers are still used to record the individual 
fragments [234a and 256b, say], while the text is recorded as "belonging 
to" accession 234a+256b. I suppose one could call that the new idno, but 
it wouldn't be quite right [and would cause lookup problems].)

But clearly the tei:msIdentifier structure (to which I should have paid 
more attention in the first place) is capable of handling all these 
possibilities.

Thanks,

Gabby

On 03/02/2011 22:26, Torsten Schassan wrote:
> Hi,
>
>> On 03/02/11 17:34, Dot Porter wrote:
>>>
>>> This does mean that you have to pick on identifier as the primary
>>> one.
>
>
> Am 03.02.2011 20:37, schrieb Matthew James Driscoll:
>> This was the whole point, in fact: you could have multiple<altIdentifier>  elements where there was no primary identifier, as in the case of "scattered" or dispersed MSS, parts of which are found in different repositories.
>
>
> Isn't it that we have to really distinguish some cases here?
>
> - If the fragments are now held together under a new shelfmark and the
> other numbers are only referrers, it would have to be msIdentifier/idno
> plus a series of altIdentifiers (cf. some remarks on this at the
> bottom.) If I understood Gabby right this is his case.
>
> - If the fragments were still separated but considered to have belonged
> to one manuscript and shall (virtually) be described as one, it would be
> the case Matthew drew upon. In this case one would use only altIdentifiers.
>
> But:
>
> - If the fragments were still sparated and if it is still valid what we
> discussed some time ago, that with TEI only *existing* things could be
> described (and therefore one could not describe something that is not
> there but only known to have been there), we would have to go for Dot's
> version: focus on one (existing) manuscript, which can be identified by
> msIdentifier and relate this ms (closely) to others. altIdentifier
> "feels" not completely right here but could be an option anyway?
>
> To follow up on that (and what I mentioned already above): In the course
> of the Europeana Regia project we worked a bit on the ENRICH ODD and we
> want to propose the following list of values for [log in to unmask] Do
> they sound reasonable, is something missing or unclear?
>
> access = accession number
> alternative = alternative writing
> catalog = number in a catalogue
> collection = a manuscript that has been grouped together with other
> manuscripts for some reason
> faulty = faulty shelfmark, but used in some literature
> former = former shelfmark
> internal = internal project identifier
> multivolume = mss is part of a multivolume and therefore has more than
> one shelfmark
> other = unspecified
> palimpsest = identifier of a previously written but deleted item
> partial = identifier of a previously distinct item
> system = former system identifier (Manuscriptorium specific)
>
>
>
> Best, Torsten
>

-- 
Dr Gabriel BODARD
(Research Associate in Digital Epigraphy)

Centre for Computing in the Humanities
King's College London
26-29 Drury Lane
London WC2B 5RL
Email: [log in to unmask]
Tel: +44 (0)20 7848 1388
Fax: +44 (0)20 7848 2980

http://www.digitalclassicist.org/
http://www.currentepigraphy.org/