[tei-council] Further ruminations on index

[Lou Burnard]

In response to Michael Beddow's comments on the proposed revision for 
the <index> element, I have now redrafted the section in question, but  
in the process, I hit on two other questions on which Council's opinions 
would be much appreciated:

a) the <divGen> element, used to mark the point where a  div typically 
containing an index or toc is to be generated, currently has no way of 
specifying headers or footers for the generated div, other than by the 
use of the global n attribute. That overloading seems to be inadequate, 
as well as potentially falling foul of all the other known problems 
inherent in text-valued attributes.  I therefore propose to change the 
content of divGen from  "empty" to  "zero-or-more elements from the 
tei.divtop class". Any objections/counter suggestions?

b) the <index> element currently has a close relative called <indexSpan> 
which is used when the indexed point is a span. The only difference 
between the two is that the latter has an additional attribute "to" 
which indicates the end of the  span. It seems to me that this is a 
rather cumbersome method of doing the job. Furthermore, since <index> 
(and therefore <indexSpan>) can now be nested recursively, the content 
model gets a little complicated. It occurs to me that we could simplify 
matters by (a) removing <indexSpan> (b) defining an attribute "scope" or 
"range" or even "target" on <index> with a default value of "here" but 
with the option to include any URL.

If you share my view that this might be a good idea, bear in mind that 
we will need to do the same thing for all the other xxxSpan elements 
(<delSpan>, <addSpan> etc.) , which is partly why I have not implemented 
it yet.  What do you think?