I don't like the <pb> - <ptr> proposal as is, because it relies on the
proximity of the XML elements to express what is basically a series of
links... and this leads me to my question: why would you not use
<link> (maybe in <linkGrp> for each source)?
Laurent
Le 29 janv. 10 à 12:49, Kevin Hawkins a écrit :
> All,
>
> Let's say you're producing a TEI edition of a text (known to us
> through various witnesses). Your starting point is the most
> authoritative critical edition previously published in print, which
> includes not only a reconstruction of the text but also a critical
> apparatus. The critical apparatus in the print version is separate
> from the reconstructed text but refers to it by line numbers and
> lemmatization, so in your TEI header, you have:
>
> <variantEncoding method="location-referenced" location="external"/>
>
> You've encoded the reconstructed text with <w> on every word, each
> with an xml:id= giving a canonical reference scheme for each word --
> for example, "W.238.16.06" for p. 238, line 16, word 6.
>
> Let's also say you will scan page images of the various witnesses to
> present with the TEI edition. You want to allow:
>
> a) a user to go from any word in the reconstructed text -- or from a
> corresponding apparatus entry -- to the appropriate page image in
> any of the witnesses
>
> b) a user to go from a particular page image in a witness to the
> part of the reconstructed text containing the words on that
> particular page image
>
> The data to do this could all be stored in the encoded critical
> edition by inserting <pb> and <cb> elements all over the place: for
> example, <pb ed="witness1" n="3r"/> to say that page 3 recto of
> witness 1 begins at this point. However, this become a nightmare to
> read and maintain. So as an alternative, I'd like to create a
> single TEI document for each witness either like this:
>
> <text>
> <body>
> <ab>
> <pb facs="manuscriptimage1.jpg"/>
> <ptr target="range(W.238.16.06,W.239.03.02)"/>
> <pb facs="manuscriptimage2.jpg"/>
> <ptr target="range(W.239.03.03,W.239.37.10)"/>
> <pb facs="manuscriptimage3.jpg"/>
> <ptr target="range(W.239.37.12,W.241.16.12)"/>
> </ab>
> </body>
> </text>
>
> or like this:
>
> <facsimile>
> <graphic url="manuscriptimage1.jpg"/>
> <ptr target="range(W.238.16.06,W.239.03.02)"/>
> <graphic url="manuscriptimage2.jpg"/>
> <ptr target="range(W.239.03.03,W.239.37.10)"/>
> <graphic url="manuscriptimage3.jpg"/>
> <ptr target="range(W.239.37.12,W.241.16.12)"/>
> </facsimile>
>
> That is, manuscriptimage1.jpg begins with word W.238.16.06 and ends
> with W.239.03.02, and manuscriptimage2.jpg begins with W.239.03.03
> and ends with W.239.37.10. Two things bother me:
>
> 1) I know of no precedent in representing information this way.
>
> 2) There's nothing to group together the <pb> and <ptr> elements (or
> <graphic> and <ptr> elements).
>
> Frankly, I would prefer if I could put a target= attribute on <pb>
> or <graphic>, but I assume this would not conform to the TEI
> Abstract Model (§23.3.3 of the Guidelines). Please correct me if
> I'm wrong.
>
> Thoughts?
>
> Kevin
|
