At 05:21 AM 7/22/2009, Serge wrote:
>I have just read through that interesting thread.
>If I summarize well there were suggestions
>to place xslt declarations at the following
>different places :
>- encodingDesc (David, Wendel)
>- encodingDesc/appInfo (James)
>- encodingDesc/tagsDecl/rendition (Sebastian, Dot)
>- profileDesc/settingDesc (Torsten)
>- @rendition of either TEI or teiCorpus (Bertrand)
>- Processing Instruction (James)
>Binding xslt stylesheets to a document is needed
>for the following purposes :
>- 'the XSLT is necessary for the "correct interpretation"
>of the document'
>- "[define] how a born digital document should be presented"
>-"[define] which xslt has to be applied to show the content
>in a certain and defined way"
>- "Recording a stylesheet seems to [be] more general than
>the view of some portions of text"
Hm. Perhaps I should have emphasized more strongly that I consider
the whole idea of asserting a normative binding between an instance
and a stylesheet to be, um, dubious.
Certainly, were I in Dot's position, I would be asking why her
questioner wanted to do this.
Not that I would regard that response as merely rhetorical. I can
think of several possible reasons that might be, given the right
context, legitimate, and there are undoubtedly more that I can't think of.
On the other hand, a number of these reasons might be answered
handily by the response "use the W3C-mandated xml-stylesheet PI to
assert this relationship", without having to invest in tagging in the
It really comes down to purposes. As I also suggested, many purposes
could similarly be served by providing the transformation results and
not the source at all.
At least construed one way, the question does seem to imply that our
normal assessment of intellectual precedence is reversed. Usually we
think of the source data as where the intellectual investment is
made, and transformation results are a useful by-product.
If, in fact, the source data is just a means to an end, to be used
only in conjunction with a stylesheet to produce a particular result,
then one needs to know when and why we preserve or publish it. (That
question isn't rhetorical either: again there may be good reasons.)
In fact, most of the time, I think both source data and
transformation results will be interesting and useful, although for
But my picking up the suggestion that encodingDesc is the right place
for recording the binding is assuming the extreme case implied by the
discussion, namely the case that takes a particular stylesheet as the
single correct "reading" of the document's encoding. I also feel this
is a heretical position, although I am certainly not above
entertaining a good heresy.
When this isn't the case (which I would hope would be more normal), I
suppose appInfo would probably be better (although the semantics of
appInfo as defined in the Guidelines would have to be revisited).
A useful counter-question to ask might be, how to encode the options
when more than one different stylesheet is mandated for use?
Wendell Piez mailto:[log in to unmask]
Mulberry Technologies, Inc. http://www.mulberrytech.com
17 West Jefferson Street Direct Phone: 301/315-9635
Suite 207 Phone: 301/315-9631
Rockville, MD 20850 Fax: 301/315-8285
Mulberry Technologies: A Consultancy Specializing in SGML and XML