[ZOOM] Possible new Connection properites - comments?

Alan Kent ajk at mds.rmit.edu.au
Fri May 21 02:37:21 CEST 2004


On Thu, May 20, 2004 at 12:15:17PM +0100, Rob Sanderson wrote:
> > Note: we are also defining some private extensions to ZeeRex in a
> > separate namespace allowing CQL relations and relation modifiers to
> > be defined. This means someone can define new operators etc in a ZeeRex
> > record *if they choose to*. If not present, we drop back to standard
> > defaults.
> 
> Could you give an example?  You can 'define' new operators using 
> configInfo/supports.  By which I mean you can record that the server 
> supports non standard operators, you /define/ them in a context set 
> document. :)

Define, declare, claim, purport, suggest, hint, lie, ... who invented English
anyway? ;-)

How about I use a phrase such as "I am using a ZeeRex record to declare
CQL index names, relations, and relation modifiers and an equivalent
attribute list that could be used in a RPN query". My CQL parser does
not care about semantic interpretations. It just cares about the legal
vocabularly of context sets, index names, relation names, and relation
modifier names; and the corresponding equivalent Z39.50 attribute list
that can be used in an RPN query which has equivalent semantics. Its
up to the author of the ZeeRex record to make sure the semantics are
the same.

> Ahh, I see your point.  We have the following query types in our ZOOM 
> implementation:
> 
> PQF    -- Obvious
> C2     -- Cheshire2 RPN representation based on CCL
> CCL    -- Obvious
> S-CCL  -- Server side CCL
> CQL    -- Translation into RPN 
> S-CQL  -- Sent through as External
> ZSQL   -- Ditto

Cool. Matthew, should the same list be added to the Java implementation
instead of the client-side/server-side flag?

Question - you say client side CCL is obvious. I assume you do the same
thing as client side CQL by default and have a fixed table of index
names your client library knows.

Alan





More information about the ZOOM mailing list