Mmmm, interesting - it also got me thinking. Essentially the information
we send on the request falls into three categories:
The query
Some positional information (startRecord, MaxRecords)
Some response formatting instructions (recordFormat, sort)
Can everything we send in a request be categorised as above?
If so, we could simplify our request into three parameters each
accepting a value which is in someway extensible.
So our requests may look something like:
http://myserver.com/sru?query=sru.CQL:dc.author=smith&positional=[sru.st
artRecord:1;sru.maximumRecords:10;myextension.stepping:2]&formatting=[sr
u.recordFormat=MODS;sru.sort=dc.author]
http://myserver.com/sru/query=sru.SCAN:dc.author%20any%20smith&positional=[sru.startRecord:1;sru.maximumRecords:10]
Any comments?
Matthew
> -----Original Message-----
> From: Ray Denenberg, Library of Congress [mailto:]
> Sent: 16 November 2007 20:29
> To:
> Subject: [search-ws] Scan: another approach
>
> Let's start over (indulge me).
>
> The cannonical scan example is:
>
>
http://myserver.com/sru?operation=scan&version=1.2&scanClause=dc.title
> =
> frog &responsePosition=1&maximumTerms=25
>
> Take away operation and version (which we're going to discard,
> hopefully)
> and you get:
>
> http://myserver.com/sru?scanClause=dc.title =
> frog&responsePosition=1&maximumTerms=25
>
> Now, we've talked about introducing multiple query types, which seems
> to be
> a well-received suggestion, and scan as one of them, perhaps not
quite
> as
> well recieved, but let's try the idea out anyway.
>
> - Suppose we call scanClause a query type (we can come up with a
> better
> name later).
> - Change responsePosition to startRecord in the above.
> - Change maximumTerms to maximumRecords in the above.
>
> You get:
>
> http://myserver.com/sru?scanClause=dc.title =
> frog&startRecord=1&maximumRecords=25
>
> Then you essentially have a searchRetrieve request. For this query
> type,
> the rule is that all records are returned via the termFormat syntax.
>
> The multiple operation and multiple database issues go away.
>
> Anything wrong with this approach.?
>
> --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
