[tei-council] classy part
lou.burnard at computing-services.oxford.ac.uk
Sat Oct 7 14:54:02 EDT 2006
Syd Bauman wrote:
> The part= attribute is defined almost identically in four places:
> * <ab>
> * <l>
> * att.divLike (<div>, <divN>, <lg>)
> * att.segLike (<c>, <cl>, <m>, <phr>, <s>, <seg>, <w>)
> The only differences are in the wording of the definition. E.g., says
> "... how the [ block | line | division | segment ] is to be ...".
> I propose
> * create new class, att.partial, which declares part=
> * add all four above to new att.partial
> * drop part= from all four, as they now have it via inheritance
This all sounds very reasonable to me.
> and also
> * make att.seLike a member of att.typed 
> * drop type= from att.segLike, as it now has it via inheritance
> * make <div> a member of att.typed
> * drop type= from <div>, as it now has it via inheritance
This sound plausible, but I'm a bit worried about the number of @type
attributes which remain not inherited -- quite a lot last time I looked.
We need to work out a rationale for deciding when this is right and when
it isn't rather than make piecemeal changes, I suggest. And your
footnote gives one reason why this particular piece maybe shouldn't be
> Someone would need to come up with good generic or exhaustive wording
> for the <desc> of part= and its values.
For "[ block | line | division | segment ] " read "element carrying this attribute" ?
>  This adds subtype= to <c>, <cl>, <m>, <phr>, <s>, and <w>. If I
> understand correctly in P4 these elements did not have subtype=
> because they were considered analogous to <seg type="phr">, e.g.
> So <seg> had the subtype=, but not the others.
Yes. So on what grounds are you now proposing to revise this principle?
I'm happy to revise it, but I'd like to know what the rationale is!
More information about the tei-council