[tei-council] containers of <relation>

Martin Holmes mholmes at uvic.ca
Tue Dec 6 13:55:19 EST 2011

I'm not 100% clear I understand this, but if <listRelation> is available 
in many places, and <relation> isn't, but <relation> is available in 
<listRelation>, is Lou's point that wherever you find yourself defining 
<relation>s it might be best to gather them together under a <listRelation>?

However, an analogy might be <linkGrp> and <link>. Both are members of 
model.global.meta, which is described thus:

"Elements in this class are typically used to hold groups of links or of 
abstract interpretations, or by provide indications of certainty etc. It 
may find be convenient to localize all metadata elements, for example to 
contain them within the same divison as the elements that they relate 
to; or to locate them all to a division of their own. They may however 
appear at any point in a TEI text."

Doesn't this explanation apply equally well to <listRelation>/<relation>?


On 11-12-06 10:33 AM, Gabriel Bodard wrote:
> Apologies, I should of course have looked at<gi>listRelation</gi>  not
> <gi>relationGrp</gi>, but the list* elements have the disadvantage of
> not being in the index next to their contents. ;-)
> My question was not about why listRelation is almost universally
> available, however, but why<gi>relation</gi>  isn't. Since we have
> approved changes base don a desire to allow @active (yuk) and @passive
> (yuk) to point to all sorts of URI-addressable elements, not just
> persons and places, why should we not by the same token make relation a
> bit more widely available?
> Lou thinks the current state of affairs is correct. Anyone else have
> thoughts?
> G
> On 2011-12-05 19:41, Lou Burnard wrote:
>> On 05/12/11 18:48, Gabriel Bodard wrote:
>>> Quick thought: we've recently redefined<relation>    to enable it to be
>>> used more broadly than originally conceived, added attributes,
>>> definition, etc., but it still is only allowed in list* elements.
>>> (relationGrp on the other hand is more widely available). Does anyone
>>> thing we should make relation available more widely as well?
>>> Why, for example, would we allow relationGrp in p, or seg, but not relation?
>> <relationGrp>   is a member of model.biblLike, as is<listRelation>   which
>> has now replaced it. This means it is available as part of
>> macro.paraContent as well as other places referencing model.bibLike in
>> their content. I am not convinced that this is a good thing.
>> <relation>   is not a member of any class, so it will only be available
>> where it is specifically named in a content model, i.e. within the
>> various list{Event|Org|Person|Nym} elements (which probably should
>> constitute a class, but don't at present). listRelation is also
>> available in those same places. This seems correct to me.

Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)

More information about the tei-council mailing list