[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [search-ws] Scan as a query type
As well as inelegance there is an implementation issue too. Basically, as an implementer, typically I'd have to handle a standard search different to a scan type search (in terms of what I search, what queries I can handle and which I reject, what record formats I'd return etc.) In the current approach, the hint as to which I should do is the different operations. In the proposed approach, I now have to pre-scan the parsed query tree looking for a scan index term before I know how to handle it. Matthew > -----Original Message----- > From: Matthew J. Dovey [mailto:m.dovey@jisc.ac.uk] > Sent: 15 November 2007 00:52 > To: Ray Denenberg, Library of Congress; search-ws@lists.oasis-open.org > Subject: RE: [search-ws] Scan as a query type > > > scan.index=dc.creator AND scan.term >= <term> > > Or > > scan.index=dc.creator AND dc.creator >= <term> > > might also work? > > > What do you think of this approach? > > I see two not catastrophic but inelegant issues with this approach: > > i) introducing scan.index into a query suddenly introduces the ability > to create meaningless queries. In CQL at present it is possible to > create silly or complicatedly obtuse queries, and it is always possible > for a server to reject a query due to it being too resource intensive > or > other implementation issues - but any syntactically valid query will > have an readily understood meaning. It is clear what the intent of the > query is, even if the server is unable to return a result set. > Scan.index introduces meaningless queries. The following simply do not > make sense: > > scan.index = dc.creator AND dc.author = "Smith" > > scan.index = dc.creator OR scan.term >= "Smith" > > NOT scan.index = dc.creator > > ii) in both Z39.50 and SRU, the query and the record format/syntax > returned are independent. If a server returns MODS, you can ask for > MODS > irrespective of what the query is. In this case, there is a dependency. > If the query includes a term from the scan set, the record syntax > returned will be a scan/index related record set. > > Matthew > > > --------------------------------------------------------------------- > 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
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]