> Date: Wed, 8 May 2002 14:21:25 -0400
> From: Ray Denenberg <[log in to unmask]>
>
> 1. Should well-known srw prefixes themselves be
> prefixed with "X-" as Mike suggests?

Hey!  That's the _opposite_ of what I suggested!

I quote:

        Date: Thu May  2 16:08:22 +0100 2002
        From: Mike Taylor <[log in to unmask]>
        To: [log in to unmask]
        Subject: Re: cql prefixes

        > I suggest that there be a convention for distinguishing
        > well-known from private prefixes, to avoid conflict (but
        > nothing quite so elaborate as what we use for Z39.50 oids).

        If we really want to go that route, then precendent (RFC 822's
        headers, the Zthes profile's Relation Types, etc.) suggest
        that something like an "X-" prefix might be used for
        extensions.
        ^^^^^^^^^^

> I don't really think that rfc 822 sets this precedent as it has alot
> of well-known fields without "x-", and it seems that the "x-" is for
> extensions.

Exactly.

> 3. What other well-known prefixes do we want?  One for Dublin Core?
> If so, what should that prefix be?  If we adopt the X- convention
> then there isn't a problem, it can be X-dc, and Ralph can use dc for
> his Dark Custard set.  Still, I'd be happier to just use dc as the
> prefix for dublin core.

Agree 100%.  Ralph can use X-DC for Deep Custard, or indeed X-Ralph-DC
if he likes.

On the matter of Dublin Core and (non-DC) cross-domain searching: it
is my conviction that the current Z39.50 XD set was a mistake, and we
would be silly to propagate that mistake in the SRW, especially given
that a large part of a reason for ZRW is to present IR concepts in
terms that will already be familiar to many people.  We should go with
vanilla DC, warts and all.

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor   <[log in to unmask]>   www.miketaylor.org.uk
)_v__/\  "It's nice to get stabbed in the front for a change" --
         Terry Venables.