Gabriel Bodard wrote:
> James Cummings a écrit :
>> Not with note/@resp but with something like origDate/@resp, you can 
>> have multiple pointers.  Unlike with persName/@ref the Guideline 
>> reference pages to seem to say what multiple pointers mean, but I 
>> would have interpreted <date resp="#E1 #E2"> as that E1 and E2 are 
>> both in some manner responsible for the determination of the element 
>> content.
> Me too. If you want to say that the resp is _either_ E1 or E2, then I 
> think we're in the territory of using <certainty/>, aren't we?

That would certainly be one way to do it.

> For the above use-case, however, I'd be in favour of standardizing @resp 
> to allow multiple pointers more widely (maybe even universally).

I certainly don't disagree. While I do agree with the others that the 
current method is really to point to a personGrp or similar, I can't 
think of good semantic reasons why: add, del, affiliation, birth, death, 
event, gap, nationality, origPlace, persName, person, sex, state, trait, 
etc. (random selection from att.editLike) allow a @resp with multiple 
pointers and many other elements allow only a single one.

Those that don't are:
att.interpLike [interp interpGrp span spanGrp]
att.textCritical [lem rdg rdgGrp]

<gap/> allows multiple @resp values but <space/> does?  So multiple 
people can be responsible for excluding a bit of text but only one can 
be responsible for noticing that there is a bit of extra space?

Dr James Cummings, Research Technologies Service, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk