[tei-council] Further purification

Martin Holmes mholmes at uvic.ca
Sat Jan 4 18:14:07 EST 2014


I see no objection to this.

I've been thinking that all attributes ought to be obtained through 
classes. I'm sure someone's suggested this many times, but it would 
solve the problem with classRef having to be able to point to an element 
to get an attribute defined on it. Where currently we define attributes 
only on elements, we could simply create a class named after the 
element. For instance, <correction> has @status and @method; these could 
reside in a class called att.correction, whose only member (initially at 
least) would be <correction>. Then if other elements needed those 
attributes, the class might be renamed to avoid confusion.

There would be the slight problem of being able to assign several 
versions of identically-named attributes (@type, for instance) to the 
same element, but I think we already have that issue, don't we?

Cheers,
Martin

On 14-01-04 02:56 PM, Lou Burnard wrote:
> Thinking about pure ODD, it occurs to me that it would be very useful if
> one could treat <classRef>s in the same way as <moduleRef>s -- i.e. by
> adding the attributes @include and @exclude to <classRef> it would be
> possible inside an elementSpec to do things like
>
> (a) <classRef key="att.global" except="att.linking"/>
>
> as a short way of suppressing a bunch of attributes
>
> or
>
> (b) <classRef key="att.linking" include="next prev corresp"/>
>
> as a neat way of saying which attributes I do want.
>
> Similarly, for model classes, it would be really nice to be able to say
> inside a content model something like
>
> (c)  <classRef key="model.choiceLike" except="seg"/>
>
> which would not interfere with seg appearing elsewhere in a content
> model, but would suppress it from inside <choice> in a tidy way.
>
> Anyone object to my adding these two attributes as part of the pure ODD
> work?
>
>
>
>


More information about the tei-council mailing list