Fw: OASIS/SRU: Scan

["Ray Denenberg, Library of Congress" <rden@loc.gov>]

We need to get the scan discussion message out tomorrow, the earlier the
better.

I've asked Matthew to draft a message, but I went ahead and started it:

----
There is discussion within the committee about representing scan as a search
function (and eliminating the scan operation).

The committee recognizes that this has been the subject of debate for nearly
20 years, though not in recent years.  Many years ago Z39.50 implementors
firmly concluded that Scan could not fit easily into the search model.
However as this has not been discussed now for several years, and as there
are new members both on the OASIS committee and in the Z39.50/SRU community,
the committee believes it is worthwhile to revisit this issue.  (The
committee does not have a position on the issue.)
-----

Matthew, if you want to pick it up from here please do so (add some
technical support for the idea, i.e. how it can be done) otherwise I'll pick
it up tomorrow.    I do think we need to initiate this thread tomorrow.





----- Original Message ----- 
From: "Ray Denenberg, Library of Congress" <rden@loc.gov>
To: "SRU (Search and Retrieve Via URL) Implementors" <ZNG@loc.gov>
Sent: Tuesday, November 13, 2007 5:39 PM
Subject: Re: OASIS/SRU: Explain/Scan


> Let's wait until tomorrow to start the Scan discussion.  We haven't put
our
> (somewhat radical) proposal out yet. We'll put it out tomorrow.  Thanks.
> --Ray
> ----- Original Message ----- 
> From: "Dr R. Sanderson" <azaroth@LIVERPOOL.AC.UK>
> To: <ZNG@LISTSERV.LOC.GOV>
> Sent: Tuesday, November 13, 2007 5:22 PM
> Subject: Re: OASIS/SRU: Explain/Scan
>
>
> > I read the thread, but don't see any mention of how you intend to
> > actually encode the scan and explain responses.
> >
> > I could see that explain could be a searchRetrieve with a single record
> > response in a given schema, but scan/browse (or 'terms' as you put it)
> > is a fundamentally different operation.
> >
> > Rob
> >
> > On Tue, 13 Nov 2007, Farrukh Najmi wrote:
> >
> > > Hi Rob,
> > >
> > > IMHO we can replace explain and scan with pre-defined indexes (or
> canonical
> > > queries) and not bloat the protocol. This also makes the search
protocol
> more
> > > RESTful.
> > >
> > > See archived thread on this:
> > >
> > > <http://www.oasis-open.org/archives/search-ws/200710/msg00046.html>
> > >
> > > BTW I have updated the issue list to fix links to archived discussion
> threads
> > > on each issue so the they can be accessed without needing to
> authenticate.
> > > The small issue list with these fixed up links are publicly available
> here:
> > >
> > > <http://wiki.oasis-open.org/search-ws/issues>
> > >
> > > Dr R. Sanderson wrote:
> > >> Yes, I do still exist...
> > >>
> > >> I have no significant objections.  It does bring SRU closer into line
> with
> > >> the way that the rest of the world seems to be going.  OAI-ORE [which
> has
> > >> been eating up a lot of my time these past few months] uses ATOM as
its
> > >> primary serialisation, so it would be nice to explore the synergies
> between
> > >> ATOM/SRU and ATOM/OAI-ORE.
> > >>
> > >> What are the intentions for Explain and Scan, which don't fit into
the
> ATOM
> > >> response quite so conveniently?
> > >>
> > >> Rob
> > >>
> > >> On Tue, 13 Nov 2007, Ray Denenberg wrote:
> > >>
> > >>> From: "Ryan Scherle" <rscherle@INDIANA.EDU>
> > >>>> Mandating Atom would have the practical effect of killing off the
SRU
> > >>>> response format,
> > >>>
> > >>> Yes but what would be the practical effect of that?  That's one of
the
> > >>> questions the committee is curious about.  How much is invested in
> that
> > >>> format, how much disruption would it cause an implementation to
> replace it
> > >>> with another, functionally equivalent format?
> > >>>
> > >>>> If it is possible to use Atom
> > >>>> in a way that feels like it was designed for search responses,
great.
> > >>> Well, ATOM is used in OpenSearch.
> > >>>
> > >>> --Ray
> > >>>
> > >
> > >
> > > -- 
> > > Regards,
> > > Farrukh Najmi
> > >
> > > Web: http://www.wellfleetsoftware.com
> > >
>