@ref doesn't currently have an example. Here's a diff
stuartyeates@stuartyeates:/opt/tei$ svn diff
Index: P5/Source/Specs/att.canonical.xml
===================================================================
--- P5/Source/Specs/att.canonical.xml (revision 12527)
+++ P5/Source/Specs/att.canonical.xml (working copy)
@@ -66,6 +66,15 @@
<datatype minOccurs="1" maxOccurs="unbounded">
<rng:ref xmlns:rng="http://relaxng.org/ns/structure/1.0" name="data.pointer"/>
</datatype>
+ <exemplum xml:lang="en">
+ <p>In <ref target="https://github.com/emory-libraries-beckctr/belfast/raw/master/xml/heaney1.xml">Heaney Workshops</ref>, the name 'Seamus Heaney' in a byline is linked to his record in the VIAF authority file using:</p>
+ <egXML xmlns="http://www.tei-c.org/ns/Examples">
+ <back xml:id="heaney1_1014">
+ <byline><name ref="http://viaf.org/viaf/109557338" type="person">Seamus Heaney</name></byline>
+ </back>
+ </egXML>
+ </exemplum>
+
<remarks>
<p>The value must point directly to one or more XML elements by means of one or more URIs,
separated by whitespace. If more than one is supplied, the implication is that the name
Hopefully the diff doesn't get mangled by the submission form.
Submission of an example (with a diff to show the suggested addition) seems uncontroversial. I think a less confusing example (even using the same link doesn't need the additional paragraph of explanation and strange use of byline in back. Just:
Marking as GREEN and assigning to me.
Example added at [r12746]
However, the example points to an html source of information which now makes the remarks provided below incorrect. (Also was already so by the example for @key.)
In general use of @ref is that people use it sometimes to point to XML element(s) and sometimes to other forms of URI. The @ref attribute provides means of locating a full definition for the entity... there shouldn't be a requirement that this definition should be an XML element.
Related
Commit: [r12746]
How about "must point to one or more XML elements or external resources by means of one or more URIs..."
That seems fairly uncontroversial.
The value must point directly to one or more addressable resources, typically XML elements, by means of one or more URIs,separated by whitespace. If more than one is supplied, the implication is that the name identifies several distinct entities.
Not sure why, but I have a mild preference for ﹤persName﹥ over ﹤name type="person"﹥.
remarks changed at [r12756] closing ticket.
Related
Commit: [r12756]