Re: [search-ws] Scan as a query type

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

Matthew -- I don't fully understand your proposed approach.  When you say
"database which only accepts searches on dc.creator", what is the term?
(i.e dc.creator=<term>)   Are you suggesting that the seed term be used? If
so, then I really do think this is semantic overload. And if not, then how
do you indicate the seed term?

--Ray


----- Original Message ----- 
From: "Matthew J. Dovey" <m.dovey@jisc.ac.uk>
To: <search-ws@lists.oasis-open.org>
Sent: Wednesday, November 14, 2007 4:13 PM
Subject: RE: [search-ws] Scan as a query type


OK, thinking this through, I believe a scan is a query on a special type
of database. However, it is a special type of database, and an
additional database.

I'll illustrate this by example:

Let is take the LoC Book Catalog - ok I can search that by dc.creator,
dc.title, dc.publisher and get back records in dc, mods and marcxml

Now a scan on dc.creator is NOT a search on the LoC Book Catalog - it is
a search on the LoC Book Catalog dc.creator index

So logically we should model the LoC Book Catalog dc.creator index as a
new database which only accepts searches on dc.creator.

Moreover dc, mods and marcxml don't really make sense for response
records when searching this database - we'd need a new record format -
let us say indexTermRecord

So to model the LoC Book Catalog with scan on creator, title *and*
publisher we would need to model this as four distinct databases (which
could be searched independently):

LoC Book Catalog searchable on creator, title, publisher, returns dc,
mods, marcxml
LoC Book Catalog Creator Index, searchable on Creator, returns
indexTermRecord
LoC Book Catalog Title Index, searchable on Title, returns
indexTermRecord
LoC Book Catalog Publisher Index, searchable on Publisher, returns
indexTermRecord

Now whilst we *could* do this, at present the SRU base endpoint
represents one database, so to do this we would need 4 endpoints. We'd
also need someway of a client being able to discover all 4 endpoints
(via explain or some convention), etc.

Or we could support a Scan operation as a special way of searching these
special index databases!

Matthew


> -----Original Message-----
> From: Ray Denenberg, Library of Congress [mailto:rden@loc.gov]
> Sent: 14 November 2007 19:03
> To: search-ws@lists.oasis-open.org
> Subject: [search-ws] Scan as a query type
>
> In light of the multiple query type consideration, I want to kick this
> idea
> around before we get into the Scan discussion on the SRU list.
>
> Is it reasonable to consider modelling scan as a query type?
>
> --Ray
>
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  You may a link to this group and all your TCs in
> OASIS
> at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php


---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in OASIS
at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php