[tei-council] vote on NH

[Conal Tuohy]

On the whole I think it could be good and useful to have, even if the
theoretical tone of the chapter is quite different to that of other
chapters. Perhaps this difference might qualify it for separate
publication, but I personally don't have a strong feeling about that.

I do think it needs more work, though. I think that the presentation of
the HORSE syntax is still problematic, in particular.

So I vote "A2", and I also second the suggestion of James that the
"Boundary Marking with Empty Elements" section be removed (or pruned
back to a stub, at least). 

The problems I see with HORSE are:

It puts up HORSE and then shoots it down as non-conformant. I don't
think I agree with this part. It seems to me only partially correct to
characterise HORSE as non-conformant. 

I am not entirely sure (my brain hurts!) but I think that some (many?)
HORSE customisations might not require any incompatible structural
changes to model classes, whose "structural properties are also a part
of the TEI abstract model" and which would therefore render the schema
non-conformant. You could argue that although a pair of HORSE elements
is semantically the same as a single regular TEI element, the difference
between using elements as containers versus using them as delimiters is
itself a deviation from the TEI Abstract Model, and therefore
non-conformant. That makes some sense to me, too, though I fear that the
whole argument is starting to sound a bit scholastic.

In any case, though, since a HORSE document can be transformed into
conformant TEI with information loss (this is true isn't it?), then it
would always be at least "algorithmically conformant" or "TEI
conformable", which is technically a kind of subclass of conformance, is
it not?

I'm not sure it's entirely correct, either, to characterise it as a TEI
extension (even though syntactically it's clearly an extension). I think
that a HORSE schema does not actually "represent concepts not present in
the TEI abstract model"; rather, it represents nothing more than the
standard set of TEI concepts, just using a different syntax. 

One other detail about HORSE, as it's presented, is that the @sID and
@eID attributes are not namespaced. Until such time as the attributes
are included in the official TEI specifications, I believe they should
be in a foreign namespace, since they are allowed to be attached to
TEI-namespaced elements.

C

On Wed, 2007-10-03 at 07:28 +0900, Christian Wittern wrote:
> On 02/10/2007, James Cummings <James.Cummings at oucs.ox.ac.uk> wrote:
> > Christian Wittern wrote:
> > > But at the moment, the question before us is NH, Yes or No?  If Yes,
> > > which of the options I outlined?  Please cast your vote.
> >
> > I suppose if I have to choose from those options, then A2.  But I think you
> > should draw the vote to a close asap, as we are running out of time.
> >
> 
> I am afraid, the time frame for the vote can't be changed now.  Given
> the fact that among the council we span pretty much all timezones and
> there are other constraints on council members time & connectivity, I
> think that 48 hrs is about the minimum for a vote.  However, if all
> votes are cast before that deadline (and I am online at that time), I
> will announce the result prior to the close of the vote.
> 
> At them moment, we have another 24hrs to go.
> 
> Christian
> 
-- 
Conal Tuohy
New Zealand Electronic Text Centre
www.nzetc.org