[tei-council] note in sourceDesc

[Lou Burnard]

There is another simple solution to the "p-in-header is different from 
p-in-text" problem out there in the wild, which I've been meaning to 
throw into the mix. This is what they do in the DTA : they have a 
different ODD for the header and for the text. Both define valid subsets 
of tei-all, of course, but each generates a different schema, which you 
can then use to validate the header and the text separately.

I've been meaning to see how effectively you could do this with a single 
ODD containing two schema specs.

I still want to put <note> inside <sourceDesc> though cos I think it 
really IS a note.

Some further comments below:


On 12/03/14 18:30, Martin Holmes wrote:
> On 14-03-12 11:14 AM, Sebastian Rahtz wrote:
>> I am unsure whether I believe in
>>
>>      a) have a system for different content models for <p> in header and in text
>> or
>>      b) making a new element for the header
>>
>> i incline to both. i.e.it is good to get a) working for when we need it, but
>> actually I don’t think its the right answer here, we need the neutral
>> “text container” element for use in the header.
> We already have <ab> for that, I think. The situation presumably is that
> people feel they are writing in paragraphs, or should be able to do so,
> not in "anonymous blocks", even when they're writing in the header.

I'm not sure that I understand this point at all. The difference between 
<p> and <ab> is not a matter of how people feel about the throes of 
composition, surely. An <ab> is just a block of text which might be 
considered prose or verse or anything. A <p> is a block of text forming 
part of a prose narrative. As I said on the ticket, there are some 
elements in the header which definitely contain this kind of <p> and 
others which arguably don't, or need not.

Note that just saying "use <ab> in the header and <p> in the text" 
though initially attractive doesn't help the fundamental problem, 
because there are definitely texts (e.g. the Bible) where the narrative 
is constructed in terms of things for which <ab> is pretty much the best 
solution
>> Allowing <note>
>> just obfuscates matters further in my book, by diluting <note>
>> to a catchall container with no real semantics at all.
> I have some sympathy with this. At the same time, I do tend to like more
> elements to be available in more places as a general principle.

I disagree that my proposed use of <note> is an obfuscation. It's an 
annotation on the content of <sourceDoc> to say "there is nothing to say 
about the source" rather than providing a bibliographic description of 
the source. <note> is for representing annotations. I suspect that much 
of Sebastian's antipathy to the idea here is fueled by anxiety about how 
to render the wretched beast.
>> BUT if we went to not having <p> in the header, and dropping DTDs,
>> we’d be in serious trouble with our users. Declaring backward incompatible
>> P6 effectively forks the TEI, and might double the maintenance.
> It's no different from moving from P4 to P5: there's a new version,
> we're mainly focused on developing that, and we do maintenance update to
> the old one for a specified period of time. Meanwhile, encouraging
> people to move away from DTDs is doing them a favour, but no-one is
> required to. I still have one P4 project that works fine, and I'm not
> planning to migrate it in the near future.

This whole DTD support issue has really NOTHING to do with the issue we 
are discussing, but just for the record, I think if we did remove 
support for DTDs, that would definitely be a move to P6. However unless 
Brussels requires us to do so, I see no need for a referendum on the 
topic. (This is an obscure reference to today's British political gossip)
>> So I vote for
>>    * doing nothing here, Lou’s <note> is just sticking plaster on a weeping sore
>>    * recommending people use Schematron to constrain <p>

See above for using <note> : is the hon member suggesting we should just 
leave weeping sores to dribble on when there's a perfectly effective 
plaister we could apply?