[tei-council] NH revised

Daniel O'Donnell daniel.odonnell at uleth.ca
Fri Oct 19 21:50:01 EDT 2007


On Sat, 2007-10-20 at 10:43 +0900, Christian Wittern wrote:
> Daniel O'Donnell wrote:
> > 1) I added a discussion of using milestones like lb to delimit
> > non-nesting text. This is in addition to other discussions in the
> > Guidelines
> >   
> This is fine, but it needs to be better integrated with those other 
> parts (eg. pointing there). 

Yes. Syd pointed this out. It is not a huge deal to fix and my rereading
now showed a couple of other first-draftisms.

> > 2) I changed the generic section delimiter from non-tei boundary to
> > tei:anchor; @type is used to indicate that the use is a delimiter;
> > @subtype used to indicate the element/feature being delimited; and @n to
> > indicate the ordinal begin/end
> >   
> This smells indeed like tag abuse.  <app> has @from and @to, which is 
> what you want here.  Unfortunately, they are not in a class -- why are 
> they not in att.global.linking or something similar?  If we could, I 
> think this is a change I would like to see.  (dare I say, there must be 
> a backdoor around those shutters...)

Syd and I were just discussing this: it is almost like we need a att.NH
model that we could invoke that came with @prev and @next and all the
other atts we use for this kind of thing. If we had one, it would also
make HORSE almost a TEI implementation.

> > 3) I dropped a section on remodelling the document class
> > 4) I collapsed the various types of segment delimitation into a single
> > section
> > 5) I added what seemed to be a missing section on join. This is in
> > addition to the discussion elsewhere in the guidelines
> >   
> that fits in pretty well.

Same as above, though: this is Dan talking about join off the top of his
head rather than after checking it out in the guidelines.

> 
> > 6) I replaced the three examples (German, Italian, and Tom Leyer) with
> > two more-or-less equivalent English ones (Wordsworth and Pinsky). I
> > wonder if the Pinsky is really better than the Leyer, however: it is
> > missing a feature the Leyer one had and replacing Pinsky with Leyer is
> > not difficult. I don't think we should have more than two main sample
> > texts.
> >   
> Right.  The examples have about the right length, whatever else you want 
> to say about them.
> 
> > The use of anchor instead of a non-tei boundary element is the most
> > controversial bit, I think. I'm concerned myself that the use of @n for
> > first and last may be attribute abuse, and Syd I know is concerned that
> > this might be abusing anchor (actually, he's sure it is; the point is up
> > for debate ;). I think I am also underselling HORSE in the current
> > version.
> >   
> We did debate the presence of HORSE here.  I am all in favor of removing 
> it for now and considering it for inclusion at a later point.

I'm more in favour of leaving it in, as it is so close to working. But
that's what this discussion is for ;)

> 
> All the best,
> 
> Christian
> 
-- 
Daniel Paul O'Donnell, PhD
Associate Professor and Chair, Department of English
Director, Digital Medievalist Project http://www.digitalmedievalist.org/
Chair and CEO, Text Encoding Initiative http://www.tei-c.org/
University of Lethbridge
Lethbridge AB T1K 3M4
Canada
Vox: +1 403 329-2378
Fax: +1 403 382-7191



More information about the tei-council mailing list