To comments of David Birnbaum and Lou Burnard on the matter of TEI
encoding of bibliography (biblScope), I would like to add the following.
In particular, with reference to:
 
(1) TEI example
 
> David says of the example 6.10.2.1,
> |where the containment of an <analytic> inside a <monogr> is
> |encoded as:
> |
> |<biblStruct>
> |  <analytic>blah blah blah</analytic>
> |  <monogr>
> |    <!-- In -->
> |    <title level="j">blah blah blah</title>
> |    blah blah blah
> |  </monogr>
> |</biblStruct>
> |
> |Is this why God created SGML comments?
 
and
 
(2) Lou's explanation:
 
> This example was included in P3 in order to demonstrate how TEI proposes
> to handle the notoriously thorny problem of the different kinds of
> relationship that can exist between an analytic entry and the various
> monographic entities which uinstantiate it . . .
 
>  [. . .] reminds me that one other possibility is
>  to use <ref>, which as a phrase level element can appear almost
>  anywhere.
 
I agree that there's a "notoriously thorny problem..." but would argue
that it's not so difficult if one focuses upon the relationships as
opposed to "sequenced" display text.  In fact -- though TEI eds may
correct me -- I think it's the TEI's emphasis upon encoding printed
texts, rather than upon storing (in this case) bibliographic
information that accounts for difficulty in using (or knowing how to
use) TEI encoding for what's been called "enumerative bibliography"
work.  More than one person familiar with BibTex or EndNote has come
to TEI and found the bibliography "encoding" model a bit perplexing.
 
In the abstract, "<ref>" would be a good candidate, since if one is
encoding the bibliographic information about the analytic level work,
*referencing* the physical host (monographic volume) to get the imprint
information would be better than redundantly storing the information
about the monographic level work every time it is "related to" the
analytic.  Well, maybe: "better" only if you have some software that
knows how to make sense of normalized data.
 
Granting all the "thorniness" that Lou references, I think it would be
an improvement to TEI, and a good candidate for new work, to focus
upon a few of the most common relationships, and to model those
relationships using some agreed-upon attributes and/or elements.
The "<ref>" (stuck in wherever you want, used any way you want)
will of course work for tagging, but that tagged text may not be
very meaningful for document interchange.
 
I don't think its fair to levy punishments for "tag abuse" in a
realm where the tag semantics are so ill defined.  Isn't this
just one example where clarified semantics for TEI could help?
 
Robin Cover
-------------------------------------------------------------------------
Robin Cover                    Email: [log in to unmask]
6634 Sarah Drive
Dallas, TX  75236  USA            >>> The SGML Web Page <<<
Tel: +1 (972) 296-1783 (h)     http://www.sil.org/sgml/sgml.html
Tel: +1 (972) 708-7346 (w)
FAX: +1 (972) 708-7380
 ========================================================================