> Date: Mon, 15 Sep 2003 13:59:37 +1000
> From: Alan Kent <[log in to unmask]>
> Being out of things a bit, I did not realise it was not possible
> already.  I think making it extensible is *very* useful. GEO is a
> good example.  But I think any Z39.50 attribute set that comes along
> that defines a new non-USE attribute will also benefit, including
> things like structure attributes (as you point out for dates).


> I understand the rationale for renaming "index sets", but "context
> sets" does not leap out with clear semantics. Got a better name?
> (CQL Schema?  CQL Profile?)

Well, if we had a better name we'd have chosen it :-)

> So I agree with Section 4, but it would be nice to have a more
> meaningful name than "context set". "CQL Profile" seems not bad -
> its even what you titled your proposal page ;-)

I'm afraid that won't do.  The things currenty called context sets are
not profiles, just as (say) the BIB-1 or Utility attribute sets are
not profiles: they're just collections of named things from which
profiles can be built.

So, for example, when we write the Zthes profile for SRW, it will
reference not only the Zthes context-set (which will provide the
thesaurus-specific index-names such as zthes.nt) but also some other
set to provide the more general indexes such as term-name,
description, etc.

This is an important distinction, as it promotes reusability of
indexes (and relations and modifiers) so that we don't find every
single CQL profile defining its own "title" index.  For much the same
reasons, in fact, that the Z39.50 Attribute Architecture uses the same

As for the title of the proposal -- "CQL - Profiling New Relations and
Modifiers" -- that reflects what the proposal is _for_, but does not
define the meaning of what a context-set is.

The obvious alternative name, in fact, would be "attribute set", since
the analogy with a Z39.50 attribute set is so strong; but I doubt
anyone wants to go that way?

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>
)_v__/\  "There is hopeful symbolism in the fact that flags do not
         wave in a vacuum" -- Arthur C. Clarke, on the moon landings.

Listen to my wife's new CD of kids' music, _Child's Play_, at