[tei-council] EDW90 proposals (1 of several)

[James Cummings]

Lou Burnard wrote:
> I'm rather slowly working my way through the proposals in Syd's table of 
> attributes in  a more or less random order, picking off the simple cases 
> first. I am simply applying the changes proposed where they seem 
> non-controversial and/or already agreed to, but this is not always the 
> case!
> 
> As the whole document is a bit long to digest, I am going to post a few 
> specific queries as they come up, to sound out Council's views:
> 
> 1. Drop the following attributes on teiHeader:
> 
> creator  (Syd proposes using resp instead)
> status (this can take values "new" and "updated")
> 
> It seems useful to me to distinguish the agency responsible for creation 
> of a header from the person who last changed it (which is identifiable 
> from the <revisionDesc>)  but maybe not?  I'm less sure about the 
> usefulness of STATUS. Anyone use these attributes?


For the record, the OTA uses @type and @date.created but none of the 
others.  I've never noticed anyone using them in other files I've 
looked at....


> 3. Drop <xxxSpan> elements in favour of HORSE-style attributes (i.e. 
> something like the spanTo attribute added to <index>)
> 
> I am basically in favour of this proposal, but it needs more careful and 
> detailed articulation

I'd agree with this (with a more detailed proposal).

> 
> 4. drop SIGIL on <witness> in favour of xml:id
> 
> This would be consistent with what we have already done for <hand> and 
> <handShift>.  But will the users of <witness> be happy having to say 
> <witness xml:id="foo"> instead of <witness sigil="foo">?

My instant reaction to this was that I disagreed.  However, I'm 
willing to be convinced.

My reason for that was I have a vague memory of using
<witness id="OneThing" sigil="AnotherThing"> previously.
My understanding was that the sigil was a nice ID which then I'd
use in my <rdg>'s etc. elsewhere.  So the <rdg> refers
back to the content of the <witness> in a semantic way as expressed by 
the guidelines, but that an id or now xml:id simply locates the 
element.  So a sigil is an xml:id with meaning.  I realise that is 
fairly tenuous.

I think I had previously mis-used them where an id was used for a 
long-form of a sigil and sigil was used for the publicly-viewable 
version.

If we follow this further, does that mean <rdg> becomes
<rdg wit="#MsA #MsB #MsC"> instead of
<rdg wit="MsA MsB MsC"> ?

If that is happening across the board with other attributes which 
function as an ID then I guess it is a good thing.  But I thought I 
should attempt to express my instant (and irrational) first impression.

-James