index-sort-as global order [was: review of index* elements]

I am about to send out an update on where I think
we stand on index* elements, so I wanted first to
send out my latest proposal on this one issue.

> -----Original Message-----
> From: Grosso, Paul [mailto:] 
> Sent: Wednesday, 2006 September 20 10:49
> To: 
> Subject: [dita] review of index* elements [was: DITA 1.1 
> Language Specification -- draft (ditaref-alpha.pdf) uploaded]

> index-sort-as
> -------------
> There remains lots of issues about "global sort order" 
> using indexterms within the prolog.  For one thing, this
> contradicts other decisions we made about what an indexterm 
> within the prolog means. 

Our current decision about indexterms in the prolog says:

 An indexterm that occurs in a topic prolog is interpreted 
 as a point reference to the start of the title of the topic.

I don't see how to reconcile this with the idea that one
can use an indexterm containing an index-sort-as in the
prolog to define some "global sort order".

There are also many details about the global sort order
concept that would need to be worked out, such as how
a global sort order within the prolog of a topic that
is referenced as the child of another topic that also
contains a global sort order gets merged and what is the
scope of the global sort order definitions.

At this point, I suggest we do away with the "global sort 
order" concept for DITA 1.1.  If and when we do define such, 
we should probably use different markup than the usual 
indexterm/index-sort-as to accomplish this.

paul