Just to add to what has already been said: the semantics which TEI addds
to xml:id really rather precludes multiple identifiers for the same
object, in my mind at least.
If we had an element <anchorGrp> of course you could use that, but we don't.
Such an extension would be easy enough to add of course (and is left as
an exercise to the reader). For the use case discussed though, I suspect
some sort of simple convention along the lines Piotr proposes would
probably be simpler.
On 11/11/11 09:51, James Cummings wrote
> I'd second what Piotr and others have said. In the conversion
> rationalize the IDs to be a single ID when they all appear at one
> place, and change any that point to those.
> There is no shortcut way to list multiple xml:id's on a single
> element. If you have multiple whitespace separated URI fragments
> in a single @ref or @target that tends to mean something specific
> to that element. Namely that all of those things together are
> identified by that element content. For example:
> <p><name ref="#Smith1 #Smith2 #Smith5">The Smiths</name> came to
> does not provide alternatives for what 'The Smiths' might mean,
> but says the phrase identifies all the people referenced. If
> this was the band 'The Smiths' I'd be pointing to a single ID of
> Whereas in something like<link> the @target attribute takes
> multiple values for a different reason, to indicate that the
> things at these IDs are somehow linked.
> I realise that none of this is what you were asking, but since we
> were talking about pointing to IDs I thought I'd mention it.
> On 11/11/11 09:32, Piotr Banski wrote:
>> Dear Steffen,
>> I think the shortest answer is that the TEI relies on an external
>> mechanism in this case (note the "xml" prefix in xml:id), and no, there
>> is no way around this, *if* you want to use this mechanism instead of
>> something hand-made.
>> But speaking of neat solutions -- you have not told us about your
>> project, but, since you mention operating on MSWord output, perhaps in
>> your processing you may be able to e.g. group all the references to a
>> single element marked with several IDs, and reduce that by throwing away
>> all but the 1st ID and redoing all the temporarily stray references to
>> use that ID alone? (Note that I assume that you group them before
>> applying the reduction step.)
>> That would eliminate the need for representing multiple IDs, and who
>> knows, maybe also the need to bother about multiple targets for your<ref>s.
>> On 10/11/11 21:31, Steffen Wolfrum wrote:
>>> as far as I have understood there is
>>> one<anchor xml:id="123"/> with one<ref target="#123"/>
>>> and there is also this:
>>> several<anchor xml:id="123"/> ...<anchor xml:id="456"/> ...<anchor xml:id="789"/> ... with one<ref target="#123 #456 #789"/>
>>> But I am looking for a way to encode the opposite in a handy way:
>>> <anchor xml:id="123"/>
>>> <anchor xml:id="456/>
>>> <anchor xml:id="123"/>
>>> with several<ref target="#123"/> ...<ref target="#456"/> ...<ref target="#789"/>
>>> Is there for this case some short version like<anchor xml:id="123 456 789"/>?!