[tei-council] [Fwd: Best way to encode bibliographic references in prose text?]
Lou's Laptop
lou.burnard at oucs.ox.ac.uk
Tue Dec 11 08:24:05 EST 2007
This does sound plausible at first blush, but bear in mind that the
<title> element really does two different things. It acts as *the* title
inside a (unique) bibliographic description, in which situtation it
definitely should not be pointing to anything else, and it is also used
to say of some string that it's a title without necessarily linking it
to anything in particular. The element for linking a string to something
else, as Conal points out in his eminently sensible reply on tei-l, is
<ref>.
So the omission of @target or @ref on <title> is not an oversight!
I can see why people might want to say that a title is special kind of
name, and should be treated in the same way as one, but I am not (yet)
convinced it's a good idea -- it seems to me to open the doors of
confusion as much as those of enlightenment.
Conal Tuohy wrote:
> Re the attached message from Markus Flatscher to TEI-L, does anyone else
> think that extending the scope of att.naming to include <title> would be
> a good idea? It seems odd to me that it doesn't already, in fact.
>
>
>
> ------------------------------------------------------------------------
>
> Subject:
> Best way to encode bibliographic references in prose text?
> From:
> Markus Flatscher <markus.flatscher at UIBK.AC.AT>
> Date:
> Mon, 10 Dec 2007 21:03:47 +0100
> To:
> TEI-L at listserv.brown.edu
>
> To:
> TEI-L at listserv.brown.edu
>
>
> Hi everyone,
>
> I'm trying to figure out the most efficient/robust/elegant way to
> encode bibliographic references in both the edited text and the
> commentary of a project in which we P5-markup most of the Nachlass of
> early-20th century Austrian philosopher Ferdinand Ebner.
>
> In medias res: Would adding a @target attribute to <gi>title</gi>
> horribly break conformity (or be likely to cause any technical problems)?
>
> I'm asking because we have lots of references to bibliographic items
> in about 20.000 pages (manuscripts and typescripts of philosophical
> works, working diaries, correspondence, etc.) in the philosophical
> estate of Ebner, and while I can comfortably address anything in the
> att.naming category, I can't figure out an elegant way to properly
> encode references to written works in the edited text.
>
> Here's an example of what I'm talking about:
>
> ===example START===
> <p xml:id="WgR.body-1.div-2.p-11">Für die Wissenschaft gibt es
> natürlich weder einen Abfall von Gott, noch den göttlichen Ursprung
> des Worts und der Sprache und so äußert sich auch schon <persName
> key="GrimmJakob">Jakob Grimm</persName> in seiner <title
> xml:id="WgR.body-1.div-2.p-11.title-1">Vorlesung über den Ursprung der
> Sprache</title> in diesem Sinne: [...]</p>
> ===exampleXML END===
>
> Note the title element in the example above. I would like for a
> transformation processor to link that <gi>title</gi> to a
> corresponding <gi>biblStruct</gi> in the back and an indexer to pick
> up a key or an id that uniquely identifies the work ("Vorlesung über
> den Ursprung der Sprache" in this case).
>
> We have a script that creates typed-path ids for any desired elements,
> and I'm aware I could link those via <gi>link</gi> (as discussed
> recently on this list), but that strikes me as brittle, especially
> since our corpus is anything but static at the moment, and I can't
> think of a feasible tracking system for tens of thousands of links
> between semantically empty IDs. If I wasn't worried about the
> brittleness, I would prefer this method at the moment for its formal
> elegance, and because it makes it possible to link any element to any
> other element. (Is there a reason why there's no global xml:href
> element, by the way? Not that that would do away with my brittleness
> concerns, but still.)
>
> I'm also aware that I could use a set of double tags of <gi>title</gi>
> (for descriptive purposes) and <gi>ref</gi> or <gi>ptr</gi> (for the
> actual link or pointer), for instance, but that, in turn, strikes me
> as inefficient; apart from that, it adds an extra markup layer on the
> element level (a [hyper]link) that is not present in the source in an
> essentially non-descriptive sense -- something I try to avoid whenever
> I can.
>
> In short, I would love to treat bibliographic references like any
> other named entity (persons, places, organizations, etc.) -- which is,
> again, why I'm wondering whether adding a target attribute to
> <gi>title</gi> or <gi>bibl</gi> would be a serious problem
> conformity-wise or lead to other potentially calamitous technical
> difficulties.
>
> In closing, I also have a feeling that I'm missing something very
> obvious here -- I'm sure there must be better solutions than the ones
> I have in mind. I'm quite clueless at the moment, and any advice would
> be extremely appreciated. Thanks everyone a lot!
>
> ===
> Markus Flatscher
> Ferdinand Ebners Gesammelte Werke (Austrian Science Fund/Innsbruck
> Brenner Archive)
> Vienna, Austria
> ------------------------------------------------------------------------
>
> _______________________________________________
> tei-council mailing list
> tei-council at lists.village.Virginia.EDU
> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
>
More information about the tei-council
mailing list