This discussion occurs fortuitously as I have finally reached a
decision on the vexed question of what to do with the <index> element
in P5. There has been some discussion of what to do about this
element, since it is a major offendor in the war against attributes,
having three of the wretched things to dispose of. Having toyed with
the various possibilities proposed and some others too, I now think
that the best course of action is to give the element a content model
where t is an element containing the text which would formerly have
appeared on a "levelN" attribute, and x is a recursively nested
occurrence of the same element.
So, if we stick with <index> and use <label> for "t", an entry which
would previously have been tagged
<index level1="fish" level2="fried"/>
The only issue on which I think debate is still necessary is the names
for the two elements concerned, since neither of these is entirely
satisfactory. To aid a speedy resolution of the matter, I have edited
the relevant section of P5 and placed a readable HTML draft thereof at
http://www.tei-c.org/Drafts/conoix.html -- comments welcomed.
In this draft, I have used <index> and <indexEntry> for the two
elements. But maybe <indexEntry> should be the outer element
(replacing <index>) and <term> (or <label>) the inner one?