[tei-council] outstanding issues in pure odd

[James Cummings]

On 03/10/14 10:54, Lou Burnard wrote:
> At the council meeting where this was last discussed,
> (http://www.tei-c.org/Activities/Council/Meetings/tcm60.xml) , I think
> we agreed on the following:
> 1.  add new <textNode> element to enable precise placing of text nodes
> within a content model, even where this is not supported by the target
> schema language
> 2.  permit @allowText on <alternate> and <sequence>
> 3.  add <mixedContent> as an alternate to <content> as syntactic sugar
> for the special case of <alternate allowText="true"> containing one or
> more <elementRef>s (though an alternate with one child seems a bit weird)

Does an alternate with allowText="true" and one child mean that 
this child is allowed to alternate with text?

I'm still a bit worried about the allowText attribute at all 
(both name and function).

> I added specs for textNode and mixedContent to the source, but nothing
> further, so these changes have not made it to the current release,
> which is annoying but gives us another chance to review them
> (preferfably rapidly) before considering some really outstanding
> questions we haven't yet addressed. I give some below in no particular order
>
> a. what about datatypes. These are currently always defined wrt either
> RNG or W3C schemas. Leave well alone?

Leave well enough alone. I think we even decided that in one of 
our face-to-faces. That we'd implement pure ODD fully and then, 
and only then, think about such other possibilities.
> b. what does <classRef allowText="true"> mean, if anything?

Hrmmm not sure. Would that mean that at this place in the content 
model you can use any members of that class but optionally 
interspersed with text?

> c. what syntax should we provide (if any) to support contextually
> varying content models e.g. <p> within <teiHeader> can contain only text
> but elsewhere must contain a sequence of <s> elements; <msDesc> is
> allowed in <text> but not in <teiHeader> We've suggested using @context
> (with an xPath) on <content> for this purpose but the details are not
> yet fully worked out.
I agree that this isn't fully worked out, but is a requirement. 
I'm not sure how to express it.

> d. some other things i can't remember
I've got some of those as well.

> As Sebastian and I are currently preparing materials for the ODD
> workshop next month (gulp), it would be good to have some degree of
> certainty about what will be forthcoming in the next release...
I'd also like to include pure ODD in the report I make about 
Council's work at the MM.

-James

-- 
Dr James Cummings, James.Cummings at it.ox.ac.uk
Academic IT Services, University of Oxford