[ZOOM] hello and first impressions of the C++ bindings

Sebastian Hammer quinn at indexdata.dk
Mon Nov 5 14:25:41 CET 2001


At 01:20 PM 11/5/2001 +0000, Robert Sanderson wrote:

>Perhaps I misrecall, but wasn't the consensus about Ashley's
>SmartConnectionBag (or whatever it was called) that it wasn't something
>we wanted to deal with?
>
>It would imply a vastly different OO model I would have thought?
>For example, you wouldn't call search on a connection object unless that
>connection object is representative of multiple connections.  At which
>point, interogating it for properties like errorCode, hostname and portnum
>becomes meaningless, so it's not a connection object any more?
>
>I'm very possibly wrong here, but that was my recollection. :)

I don't recall such a conclusion... But look at Mike's Perl "binding" 
(quotation marks because it may need revision to track ZOOM), which does 
support asynchronous operations against multiple targets, or look at Adam's 
"C" binding which should have a slightly different take on the same thing.

I don't know about you guys, but 90% of the clients I build for customers 
have to support parallel searching.. it's one of the primary selling points 
of Z39.50 in libraries. Yes, you could do it serially with a single-target 
ZOOM binding, but at a prety severe performance penalty.

To what extent the OO model holds up *is* a reasonable issue to discuss, 
though.

--Sebastian




More information about the ZOOM mailing list