> Date: Wed, 17 Dec 2003 10:30:36 +0000
> From: Robert Sanderson <[log in to unmask]>
> > Really?!  Reference, please.  I thought I was quite at liberty to say
> >         >cql="" cql.title=dinosaur

You mean the bit that says "The prefix of 'cql' is reserved for the
default CQL context set, but authors may wish to recommend a short
name for use with their set."  If I'd spotted that before, I'd have
called it.  I don't see any reason to special-case this.  If I want to
map the "cql" prefix to my private set and have the meanings of all
the relation modifiers change, that's surely my decision.

> In ZeeRex, I would say:
> <default type="cql">1.1</default>
> <supports type="cql">1.0</supports>
> And then make it mandatory to include the CQL set if you support it.
>   <set identifier="..." name="cql"/>


> And by extension, making it also mandatory to declare the DC schema
> if supported, a reversal of what we said earlier.
>   <schema name="dc" .../>

This is a separate discussion, right?  Or have I missed a connection?

> > CQL Context Set
> > Record Metadata Context Set
> The context set will live at the LC site, where its identifier
> is. And (as yesterday) the site will still exist as
> there's no reason to get rid of the host name, but I'd prefer
> The Identifier will definitely change.

Why not put the RM set on the LC site, too?  It is very much part of
the core group of context-sets, along with CQL, the Network set and
DC.  I'd prefer to see them all live on the LC (unlike clearly
application-specific sets such as my Zthes and your CCG).  This is
especially true after the OCLC debacle where the URL of the
Cross-Domain Attribute Set changed.

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>
)_v__/\  "In order to progress as a writer you must learn to struggle
         to put proper words in proper places" -- Bret Jolly.

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