Poul Henrik Jørgensen wrote:

> I therefore support Ralph's views below.

Ralph's view seems a bit terse:


> From: LeVan,Ralph [mailto:[log in to unmask]]
>
> I'd be happier with the extensible XML document than an interface that might
> need to change.

And further...

Alan Kent wrote:

> Put another way, a ZeeRex record may be able to describe a SRW server.
> But does that mean SRW should use a ZeeRex record as the SRW explain
> mechanism?

I'm trying to understand what the three of you (Ralph, Poul Henrik, and Alan) are
saying, and I think it's this: You would rather see the burden placed on
client-side processing; thus,  send all the explain information and let the
client extract what it wants.

I agree, and this was the premise of srw/explain from the start, that there would
be a single explain record; we wouldn't be searching for and
extracting/retrieving specific explain information.  I don't see how this is
fundamentally inconsisitent with ZeeRex though.  A ZeeRex database will have
records each describing a single database and for normal Z39.50 you will search
it.  But an srw server would correspond to one database, and so would have a
single ZeeRex record.

I also sense that you (Alan in particular) aren't comfortable that ZeeRex meets
the needs of srw.   So I think it's time that we determine what these
requirements are, before we can have much more meaningful discussion on this.
Could you three please give some thought and list what you think are the ZeeRex
requirements? Thanks.

--Ray