At 2001-04-16 08:16, Tobin Nellhaus wrote:

 > So my main question is not so much whether TEI tags can be adapted
 > to fit the need -- by adding a tag set or in some other way
 > modifying the DTD, they certainly can -- and more whether TEI has
 > adopted the most suitable philosophical assumption in establishing
 > the existing drama tag set.  Perhaps an analogy within TEI would
 > help.  In theory the verse tag set could do without the line <l> and
 > line group <lg> tags, and instead use various combinations of
 > (slightly adjusted) <p> and <divn> tags -- but would you want to?
 > Imagine the conceptual contortions one would have to go through if
 > those tags were missing.  Imagine the protests from encoders of
 > verse!  It was my understanding that, even though encoders often
 > modify DTDs, the tag sets are meant to cover the concepts of
 > structure that are most basic to each genre (plus a few that are
 > more unusual), and that's for reason the plethora of tags.

The concepts which are most basic, or the concepts which are most
clearly agreed on.  Remember that 'basic' does not necessarily mean
'most fundamental'; it may mean only 'unavoidable'.  You suggest that
the TEI drama tag set is better suited for describing the text than
the performance; I think that's probably a fair assessment of the DTD
without any extensions.

There are several plausible reasons one might assign for this:

   - it's relatively clear how to represent the text, and the
     tradeoffs among choices are reasonably well understood; it is much
     less clear and there was much less consensus about how to
     represent aspects of a particular performance, or even what
     aspects of a performance are most often likely to be recorded by
     those with an interest in performance (blocking? sentence
     intonation? length of pauses?  lighting cues? synchronization with
     a videotape or audio recording?  contents of the stage manager's
     copy of the script?  lighting booth copy? director's copy? ...)

   - the best way to represent a performance in linear forms depends
     on what you want to do with the representation

   - those interested in the drama (which I take as the name of a
     literary genre) need the text, not the performance, while those
     interested in the performance will often (perhaps usually) need
     both a representation of the text and a representation of the
     performance

   - the TEI currently does best with written artifacts

   - the TEI currently does best with abstractions of written
     artifacts (it does less well with the physical description
     of text carriers)

There are also some implausible reasons (by 'implausible' I mean I
think they are false):

   - the text is more important than the performance

   - the text is not intrinsically more important than the
     performance, but the creators of the TEI tag set for drama thought
     that it was

   - the creators of the TEI wanted to discourage the study of
     theatrical performance

   - the creators of the TEI thought written artifacts were the
     only important thing, or that linguistic objects were the
     only important thing

You may well be right that some fundamental change of assumptions
would be the most helpful way of developing a way to represent
performance information.  Even if you are not right about this in the
general case, you may well be right in feeling that the assumptions
made in the TEI drama tag set are not congenial to your point of view,
or even that they are in flat contradiction to it -- in this case it
is not only your prerogative, but also (some would say) your
intellectual responsibility to design a tag set that reflects the
assumptions you believe are fundamental.  And I believe anyone who
regularly reads this list will be very interested to see what you come
up with. (And, for what it is worth, assuming that you document your
DTD properly and provide a header, you can claim that the result is
TEI-conformant.)

I would suggest, however, that any attempt to represent a performance
which does not manage to represent the script is likely to be
problematic for many purposes.  It is in this sense that the tags in
the drama tag set represent 'basic' information.

 > In other words: *Can* these documents be marked up just by
 > using/adding existing tags? Yes, although with a bit of difficulty
 > and possibly with some maltreatment of the TEI guidelines.  But
 > *should* it be done that way?  I don't think so -- not when looking
 > at something that is a normal and *essential* part of drama, even if
 > completely unrepresented in the ordinary published script.

I'm not sure I understand what you mean -- prompt books and so on are
a normal part of modern Western European theatrical practice (and
possibly of theatrical practice in other cultures as well), so I
suspect you are using "drama" here as a synonym for "theatrical
performance" or even "literary works destined for theatrical
performance".  The term "drama" does get used more broadly, and drama
(qua literary genre) can certainly exist without performance and thus
without prompt books.  The absence of prompt books for any performance
of Tolkien's verse play about the battle of Maldon does not seem to
mean that it is not drama.  It does means the play has never been
performed. The absence of prompt books for the few chapters of Moby
Dick which are written in dramatic form doesn't mean that they aren't
in dramatic form; it only means they aren't performed or (in the usual
sense of the word) performable.

But my reservations about your statement are a side issue.  You seem
to be working yourself up to the conclusion that tagging prompt books
might require you to add new elements and attributes to the DTD, and
you seem (a) to be unhappy with the conclusion and (b) to believe that
it is best motivated by the view that the TEI has missed something
essential.  I think you are making heavier weather of this than you
need to. It can be desirable or even necessary to extend the TEI DTD,
even without the DTD having missed something essential.  (In the Model
Editions Partnership, for example, the 'hi' element is supplemented by
specific elements for underlining, italics, bold, and small caps.
Nothing essential was missing from the TEI DTD, but having distinct
elements helped make the MEP data capture DTD somewhat more convenient
for editors.)

 > (Likewise, not all poems use stanzas or have line segments, but the
 > <lg> and <seg> tags are considered a normative part of the verse
 > set.)  Bear in mind that I'm suggesting the need for only a handful
 > of new tags, which is not much in the TEI world.

As Syd Bauman has succinctly said, there is nothing wrong with
extending the TEI DTD.  The DTD is designed to be extended, and a fair
bit of work went into the design of the extension mechanisms.  As Syd
points out, even TEI P3 uses the extension mechanism -- in part to
underscore the idea that there is nothing wrong with extending the
DTD.

You may, if you wish, harbor the thought that if the designers of the
DTD had only done their jobs RIGHT, no extensions would be necessary,
and thus that any extension identifies a place where a mistake has
been made, but you will have a hard time persuading some of us.
Opportunities for improvement in a future revision, yes.  Mistakes,
maybe but not necessarily. One of the key axioms of the TEI design is
that "No markup language with a finite vocabulary can be complete" in
the sense of handling equally well all texts from all periods and
languages, for all different scholarly or non-scholarly purposes.  It
follows as a consequence that "in order to be generally useful a
markup language must be extensible." (L&LC 6 (1991): 36).

If I were tagging stage managers' books as I was taught to make them,
I would almost certainly add specialized element types for lighting
cues and sound cues and "WARN RAFI" and the like.  But I would be
puzzled by any suggestion that this would tell us much about any
particular *performance* of the play.  The prompt book, alas, does not
tell us whether the seagull cries actually came in at the right
moment, or after too long a pause.  (And the note "WARN RAFI" becomes
relevant to the audience's perception of the performance only when the
stage manager fails to perform the task and Mrs. Rafi misses her
entrance.)  At best the prompt book tells us about a production, not
about a performance.  And it's quite a narrow view at that.

As you foresee, and as Wendell Piez has confirmed, you are going to
face quite a challenge in handling synchronization (light cues that
cover three and half speeches) and dramaturgical revision of the text.
Some of the techniques in Barnard et al. 1995 may come in handy; those
in the chapters on synchronization and alignment and manuscript
transcription may also be useful.  Some specialized elements may make
the tagging more convenient.  But overlap is hard to tag in SGML
and XML in ways that are both convenient and easily processable.

Good luck.

-C. M. Sperberg-McQueen