A few days ago, Michael Sperberg-McQueen wrote against giving
presentational specifications for the TEI-specified document structure,
saying among other things,
> ( . . . ) and it would distract our slender
>resources from the crucial and difficult specification of the tag set
>and structure into an enterprise which, however useful, does not pose
>any *intrinsic conceptual difficulties*. . . . (emphasis added)
Granting the weight of other reasons for the TEI's decision, the
emphasized point raises such a red flag for me as to call for
reconsidering the whole question. A lot of us in the computing business
have burned ourselves and others with similar arguments, taking the line
that because there are no known conceptual problems with an enterprise
it will be easy in practice. In the black humor of the software
business, this is described as an SMOP -- "simple matter of
programming". Projects have frequently failed because the time and
difficulty of the SMOPs were grossly underestimated. The damage in
these cases goes beyond the obvious schedule slippage and cost overrun.
Since the extra time and effort go into work initially defined as
'easy', managers' instincts say the time is wasted and shows
incompetence; programmers' instincts say the time can't be necessary,
and push them to doing a hasty job.
In the case of the TEI, the very sound arguments against specifying
presentation face quite a clamour for *some* presentational form that is
more human-readable than is the SGML-tagged text stream. I suggest that
the TEI should define a simple but readable presentational interpre-
tation of the standard, to be called "TEI examination form", explicitly
*not* standard for publication or formal presentation. TEI examination
form should include some documented variations, e.g. for presentation of
diacriticals and emphatic markup on devices that do not support such.
It would then be *one* presentation that TEI-supporting software should
support, and that would make sense to all TEI-knowledgeable users. The
alternative may be much reduced usefulness for the standard, with
marked-up text little used because of its inaccessibility to humans, and
formatted text having proprietary forms so that the unifying effect of
the standard is lost.
This does not mean TEI should go into software development. The "TEI
examination form" is proposed as a set of presentation rules, not as
code implementing those rules. Nor does it mean that every TEI tag
should have a presentational form; it means that enough should have such
a form that the remaining tags retained literally in the text are not a
bar to human reading.
Re: Do presentational specifications belong in TEI?
Sounds reasonable to me.
The basic notion would seem to be that it should be possible to
specify a set of presentation markup capabilities that are
nearly universal, such as line break, indent, center, flush left/right
footnote, etc. and then allow/require all TEI tags to have some
specifications in these universals. It might even be desirable to
have a more elaborate set, including bold, italic, etc. and allow
those as well to be acceptable accompanying traits of the tags.
This alternate formatting language could be made reasonably
device independent by preferably not assigning any numbers to
anything, making it possible to realize this concrete presentation
format on many devices of varied output.
I still also believe that SGML itself should worry about its
presentation format when written down with tags on a page.
I've been informed by professional publishers that they
experienced considerable problems due to SGML's failure to
exactly specify the nature of blanks and line breaks around tags.