[tei-council] encoding page scans

[Conal Tuohy]

Afternoon all! My first post on the TEI Council list is about encoding
images of page scans in TEI. 

A couple of months ago Jamie Norrish started a short thread on TEI-L
about this issue, but I don't feel that the issue has been adequately
addressed. Of course, it may be that I've missed something in P5 and
that this is now all sorted. I hope so :-)

If I haven't missed anthing, then at the moment there isn't a really
good way to encode these scans (they are not figures, but graphics
having a different relationship to the text). Many people are encoding
these in TEI, but there's no consensus on how to do it. Recently I've
come across another TEI customisation in which the URIs of scanned pages
are encoded as pb/@url. I've seen others in which it is encoded as
pb/@entity (a reference to an unparsed entity, like P4's
figure/@entity). There are other practices in use too. 

At the NZETC we are encoding page scans (in P4) using a list of figure
elements inside a note in the TEI.2/teiHeader/fileDesc/noteStmt, each of
which is linked to the corresponding pb element with a @corresp. It
seems awkward (to put it mildly) to encode the figures in a notesStmt,
but it seemed preferable to encoding them as part of the text (as had
been our practice). 

Of course METS provides one solution. But it seems like overkill to add
a METS wrapper around a TEI file just to support a single feature which
(to me) seems already like a natural piece of TEI, and which has several
times been hacked into TEI anyway. In the earlier thread, Christian
argued that the images were alternate representations of the whole text,
and hence didn't belong in the file at all. I can see his point (I
think), but I don't find it convincing. 

Can we establish a standard mechanism in TEI for this purpose? Off the
top of my head I would prefer to nest a graphic element inside a pb, but
to me the important thing is just that some simple mechanism is provided
within the standard TEI schema without customisation.

Cheers

Con

PS incidentally, I note that graphic/@scale has type data.probability.
Shouldn't this be a real number? 

--
Conal Tuohy
Senior Programmer
+64-4-463-6844
+64-21-237-2498
conal at nzetc.org
New Zealand Electronic Text Centre
www.nzetc.org