[tei-council] Pure ODDness

[Martin Holmes]

On 13-11-20 03:27 AM, Lou Burnard wrote:
> As I proceed with implementing Pure ODD, inevitably a few
> obscurities/uncertainties surface. Advice from Council members would be
> much appreciated.
>
> 1. un-mixed content models
>
> We decided to represent mixed content by means of an attribute @mixed on
> the relevant container element (<sequence> or <alternate>). But how do I
> represent a content model which in RelaxNG would be just plain old
> rng:text ?
>
> The choices seem to be
> a) also allow @mixed on <content> and say <content mixed="true"> (even
> though actually the content isn't mixed at all)
> b) change the name of @mixed to something like @hasText, @allowsText,
> @textual  and go with option (a)
> c) introduce a model.contentPart element called <pcdata> and say
> <content><pcdata></content> (or some better name)
>
> Preferences? Alternatives?

I go for c); it's parallel to RelaxNG and will be easier to understand.

>
> 2. co-existence
>
> I think we agreed that the old and the new method of defining content
> should co-exist at least for the next release. How should that be done?
>
> a) change the content model of <content> to permit either
> model.contentPart  or macro.schemaPattern
> b) change macro.schemaPattern to permit either model.contentPart or
> macro.anyXML
> c) define a new element (howabout <model>) containing model.contentPart
> ; allow both that and current <content> initially, then deprecate <content>

c) seems cleaner to me.

Cheers,
Martin

>
> Now that I've enumerated them, I think I know which options I prefer,
> but it would be good to see whether anyone agrees.
>

-- 
Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)