[tei-council] numbered divs, a proposed solution

Mon Jan 8 19:09:13 EST 2007

Syd Bauman wrote:
>> the ingenious hack #1 only works for DTDs. 
>>     
>
> Why do you say that? Both the .rng and .rnc files I get from roma are
> valid, and seem to work exactly as expected (i.e., as the DTD does).
>   
yes, sorry, I lied. The dummies are not needed in the
other languages, of course.
>   
>> But it puzzles me as to _why_ it works, since the content model
>> _is_ ambiguous.
>>     
>
> Again, what makes you say that? The content model for <body>, e.g., 
> certainly looks deterministic to me.[1] Xerces J, xmllint, and
> onsgmls all say the DTD is fine.
>   
yes, its technically OK. but if you have

 (DUMMY | foo), (DUMMY2 | foo)

and you meet just a "foo", where does it come from?
>
> Allow me to play devil's advocate. First, I will argue that it's not
> an at all difficult task. To remove numbered divs go to the sample
> ODD, find the part that says "copy-and-paste this section", copy it,
> paste it into your ODD, and you're done. Heck, there could even be a
> little check-box on the Roma web interface that caused this
> oh-so-common request to happen for you.
>   
I wish I could go along with saying that's acceptable, but I just
can't.....

> But more to the point, when did it become the TEI's goal to protect
> users from schema code? 
speaking for myself, since the dawn of creation. its axiomatic
in my book. I realize we have different books :-}

Anyway, having divLike classes adds the enormous advantage that
you can now easily add your own new object to be a child of <body>.
Don't tell me *that* was easy before!

-- 
Sebastian Rahtz      

Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

OSS Watch: JISC Open Source Advisory Service
http://www.oss-watch.ac.uk