[Ex-plain] RE: Explain--
mike at tecc.co.uk
Wed Apr 3 17:47:56 CEST 2002
> Date: Wed, 3 Apr 2002 10:22:53 -0500
> From: "LeVan,Ralph" <levan at oclc.org>
> Thanks everyone for taking this stuff seriously!
No problem -- it's serious stuff!
> I think the topic of general local extensibility is a good one. If
> I am to replace my current configuration files with this, then I'm
> going to need stuff that you won't want. The location of a holdings
> database is a good one, though I think you'll hear more calls for it
> when the topic leaks off this list and onto the Holdings list.
I'm sure you're right -- that's why I questioned whether the term
"extension" fully captures the concept that we're edging towards here.
I think we'll find that we want to support BOTH community-specific
extensions (e.g. location of holdings database for the biblio
community, BSI or ISO-conformant for the thesaurus community) AND
purely local extensions, like OCLC Unique Database Id for you people.
Probably the same extension _mechanism_ is right for both of these,
but there will be social engineering issues to negotiate too
... forwarned is forearmed.
> Mike, I'm afraid that "blows up" does mean catastrophically. We've
> got code to react gracefully to a diagnostic. But, some servers
> just drop connections.
Hmm. Well, we all know that the Correct (capital "C") solution to
such servers is just to ignore them; but I do recognise that this is
not always the right (lower-case "r") thing to do :-)
> The NamedResultSets option bit didn't show up until version 3 and is
> not supported by the servers that blow up when you send them a
> result set name other than "default".
Er. So named result sets were in v2 but there was no way to negotiate
them, is that it?
(And do we need an Explain-- element to specify v2/v3?)
> I vaguely remember a case of a server that ignored result set names
> and we turn this on for them as well.
We could question whether it's really Explain--'s job to step in and
correct the lies told by so fundamentally broken a server, but I guess
this is another case where pragma wins out over purity?
> Thanks, again!
Thanks for getting involved. We're keen to make sure this meets
people's real needs rather than just things we think they _should_
need, so this sort of feed is crucial.
/o ) \/ Mike Taylor <mike at miketaylor.org.uk> www.miketaylor.org.uk
)_v__/\ "You cannot really appreciate Dilbert unless you've read it
in the original Klingon." -- Klingon Programming Mantra
More information about the Ex-plain