Re: position conditional patterns

Bob Foster wrote:

> The design does seem a little peculiar. The sort keys are actually keys 
> on refType. Why doesn't the refType element specify both the primary and 
> alternate sort keys and get this order-dependence out?

That might indeed be the way to go. I wonder if you have a suggestion on 
a related issue?  In some cases, the alternate sort key would map to an 
element, but in other cases it would just be text.

For example:

<reftype name="book" sort-key="creator" alternate-sort-key="Anonymous">

vs.

<reftype name="article" sort-key="creator" 
alternate-sort-key="container-title">

So in the first cases basically your validating again "text" while the 
second would be a choice of strings.

Any good ideas on how to handle this?

Bruce