[tei-council] Proposal <idno> coverage -SF 2493417

Lou Burnard lou.burnard at oucs.ox.ac.uk
Fri Jan 23 04:35:50 EST 2009


The big big difference between the docbook view of an author and the 
TEI's is that the TEI model, rightly or wrongly, wants to allow you to 
just put plain old text in there, which docbook doesn't permit. So you 
can't say (in Docbook) <author>Unknown</author>, much less <author>La 
Fayette, Marie Madeleine Pioche de la Vergne, comtesse de 
(1634–1693)</author> (and, while comparisons are odorous, I note that 
the docbook model also doesnt permit you to do much about that 
"(1634-1693)" which many librarians insist on)

Changing that possibility would be a fairly fundamental design decision. 
I don't say a bad one, but certainly one guaranteed to break more than 
80% of the world's known TEI documents and thus proscribed by the 
Birnbaum doctrine.

It seems to me that one way out of this argument might be to propose a 
new <authorStruct> element, which doesn't permit text, but does require 
the kinds of elements we're all agreed form a sensible part of a modern 
bibliographic structure. Strange that the library community hasn't been 
asking for this all along really. Maybe we should move the discussion to 
TEI-L to see if this would fly?

As to the bizarre things that crop up in the TEI's current content model 
(which is model.phraseSeq) -- that's because we never really finished 
the job of defining model.headerRestrictedText -- i.e. a subset of 
phrase level elements which make sense as header content.


David Sewell wrote:
> On Thu, 22 Jan 2009, Laurent Romary wrote:
> 
>> You should not have some prejudice as to what scenario a user has in
>> mind. We need to provide both ways.  When a bibliographical reference
>> comes in the header of a scientific paper or as an export from a
>> publication archive, you don't want to reconstruct the information by
>> following links here and there. You need self-contained <biblStruct>'s
> 
> I would agree with Laurent here. If it is the case that standard author
> identifiers are becoming part of the scholarly bibliographic world,
> allowing <idno> as a child of <author> makes sense, particularly for
> newer users of TEI encoding for whom the use of @xml:id / @corresp 
> referencing schemes will be somewhat confusing. (And by "confusing" I 
> don't mean the idea of pointers, rather the appropriate or legal places 
> where referenced elements may appear in a TEI document.)
> 
> The proposal would be somewhat analogous to the use of <uri> in the 
> DocBook content model for <author>:
> 
> http://www.docbook.org/tdg5/en/html/author.html
> 
> which can be used to provide an identifier:
> 
>    <uri type="nldai">info:eu-repo/dai/nl/12456454</uri>
> 
> (and as an aside, if I were a beginner trying to decide whether to use 
> TEI or DocBook for encoding an original article, I would find the 
> DocBook <author> element reference much more comforting than
> 
> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-author.html
> 
> with over 100 legal child elements, most of which have no apparently 
> connection to the concept of an "author".)
> 



More information about the tei-council mailing list