[tei-council] Tite ticket - pb at facs

James Cummings James.Cummings at oucs.ox.ac.uk
Thu Feb 24 08:48:57 EST 2011

On 24/02/11 12:31, Lou Burnard wrote:
> I've no problem with the idea of saying that only<pb>  should get @facs
> in lite, but the point of my comment was that it is not obvious how to
> avoid other things getting it too. In general, I think "cloning"
> attributes is not good practice, so it would be a useful exercise to see
> whether it is now possible to make just one element in a schema inherit
> from a supposedly global class. That was after all one of the purposes
> of the changes made for release 1.9

I've made a couple trial-and-error tests on this but everything I 
try so far ends up giving @facs as well to the Tite-specific 
elements like <ul> ...  I must be missing something though.  I 
think it is because if I include att.global.facs then everything 
that is specified in the ODD as being changed to get att.global 
(i.e. new elements) but this doesn't seem to happen with the 
unmodified elements.  I've probably just confused myself. 
Suggestions on how this *should* work?

> In response to Kevin's comment on the ticket: there is a big difference
> between pb at facs and graphic at url -- the latter means that there is an
> image here which forms *part of the text* -- the former means that there
> is an image of this part of the text here, but it is not a textual
> constituent.

Seems entirely clear to me.

> I'm also a bit surprised to learn that data capture agencies want to use
> this feature: are they going to be supplying page images too? If so,
> shouldn't we be going the whole hog and adding<facsimile>  and<surface>
> (at least) as well?

I suppose this is treading the borderline between having a very 
tight Tite and adding in useful information.  Most scanning and 
pagedescription formats that produce XML output include some sort 
of coordinate systems for the accompanying page images. (I'm 
thinking here of things like Omnipage XML and DejavuXML both of 
which I've written bespoke translations to TEI for.)  So having 
things like surface/zone seems entirely in keeping with the needs 
of mass digitizers I would have thought.  But that said, if you 
gave me a <pb facs="foo.jpg"/> the automatic creation of 
<surface/> elements from that is fairly scriptable as there are 
tools that will give one the dimensions of the image easily.


Dr James Cummings, InfoDev,
Computing Services, University of Oxford

More information about the tei-council mailing list