P5 pointing needs

[Lou Burnard]

This is the note which Syd mentioned during our discussion of the lack 
of progress from the current SO group. Could council members please 
review this plan and comment as to whether they are broadly in agreement 
that (a) this is an acceptable outcome from the group (b) the proposed 
timetable is acceptable?

<p>-------- Original Message --------
Subject: proverbial whip ... P5 needs to point
Date: Sun, 28 Mar 2004 12:18:22 -0500
From: Syd Bauman <Syd_Bauman at Brown.edu>
Reply-To: Syd_Bauman at Brown.edu
To: David G Durand <David_Durand at Brown.edu>,        Lou Burnard 
<lou.burnard at computing-services.oxford.ac.uk>

Council conference call will be Tue morning 08:00.

As I see it, the only absolute high priority thing to do is to get to
us a coherent system for pointing. Figures, SVG, stand-off mark up
linguistic examples, canonical references, etc., are all important,
but none of them is holding up other work. By this point the lack of
a pointing policy is in fact causing (future) delays in P5. We
(actually, Lou gets all the credit) are hard at work writing new
chapters for P5, and in many cases need to know how to point from one
element to another. What is being written now may have to change, and
minimizing that seems pretty important.

IMHO what we need from the WG ASAP is a general sketch of how
pointing (and thus linking, joining, etc.) will work. This material
is mostly already covered in SOW02 and the last section of SOW08,
although the former is of course vague and offers choices where we
actually want solutions. The only things that I think are missing are
* what to do about the identification of elements (do XPointers point
   to id=, name=, any ID attribute, or what?);
* what to do about IDREFS attributes (WG discussed this, but I don't
   remember when, or what if any resolution was reached);
* what to do about pointing to spans or ranges (where I think the
   answer may well have to be "we can't for now", or some horrible new
   indirection element soley for this purpose).

Lou may different thoughts -- I have not talked with Lou about this
for a long time.

Also need to tackle the question of element-by-element or
document-by-document choice of pointing mechanisms. On this one I've
sort of developed an opinion: TEI schema should permit
element-by-element choice; additional modules should limit this choice
for entire documents

So I see two really quite important steps.

1) In the almost immediate future the WG needs to produce
    instructions on pointing so other WGs can get on with P5. While it
    would certainly look best to have it before the conference call, I
    think it is unlikely to make any significant difference (at least
    to me) for a few days thereafter. So I think Tue 06 Apr should be
    a deadline for this.

2) A full draft of the replacement chapter 14 should be *finished* by
    Thu 06 May.

If you honestly think these deadlines are going to be met, well and
good, let's get on with it. (Again, I'm happy to help.) On the other
hand, if you feel there is any significant chance they can *not* be
met by the WG as currently construed, I think we need to start
discussing alternative arrangements for getting this material drafted
to a useable state. (Not that I have any idea what, if any, process
there is for such arrangements.)

Documents on graphics (and pointing thereinto), stand-off-markup,
linguistic examples, etc., could probably trickle in over the summer
without any significant detriment to the P5 process.

Feel free to call me. I'll be home most of today and tomorrow
morning; tomorrow afternoon is no good due to dentist and physical
therapy appointments.