For much of this debate over @facs I have been inclined towards Martin's
point of view. This is partially out of sympathy for anyone who asks a
seemingly simple question and gets a discussion of and even criticized over
their choices regarding several other subjects, but it is largely because I
operate from the perspective that the Guidelines as published must be
assumed to be correct and complete as written (until changed) and should not
be changed without a very good reason. After all, we must presume, for the
purposes of preserving compatibility, that some users who do not read this list
have already planned out and are creating their documents and/or applications
based on their interpretations of the P5 1.0 Guidelines as presently written
and may not be particularly open to or even notice a late change.
That being said, I must admit that I find convincing Con's argument that the
TEI's plethora of linking/pointing attributes, which gave me a bit of a
headache when I first puzzled my way through them in P4, (should) exist to
allow machine-interpretable distinctions between types of relationships. I do
see making machine-readable such human-created interpretation--of the
digital text itself as well as of any base text--as part of the point of TEI.
More experienced philosophers should feel free to correct me on that if I'm
wrong, of course.
However, Con's argument starts with a view of one of TEI's underlying
philosophies and works its way forwards from there. If we are to truly have
this debate, it would be best to temporarily table the questions of specific use
cases (OCR, IMT) and even of specific attributes (@facs, @corresp) and start
with philosophy. Here is my question sequence:
#1: Do we agree with the argument that the underlying philosophy of our
pointing/linking systems is to provide separate attributes (and elements?) for
separate types of relationships?
#2: If so, what relationship(s) is(/are) each attribute for? Are there "not-
otherwise-specified" or catch-all attributes? (More than one would suggest
levels of generality; one might catch everything in a specific category, while
another would catch everything that didn't fit a category.)
#3: Does theory (#1) match practice (#2)? Is there anything that can be
done to make them match better?
#4: When would it be appropriate to make the changes? (I.e., are these just
clarifications to P5 1.0? Are they P5 1.1? Are they P5 2.0? Are they P6?) Is
such a change going to cause compatibility problems?
Some of this may be more appropriate for a more formal change process; I
haven't been around long enough to know precisely how these things work.
Regardless, if we are going to base our arguments on the philosophies
underlying TEI, we need to make sure that we agree on those first.
M. Alan Thomas II
P.S.: Discussions like this are one of the reasons why I never feel inclined to
just throw out unread list messages when I come back after a few days away
from my e-mail. Thanks for keeping things interesting.