- From: Alison Stevenson <as@it-innovation.soton.ac.uk>
- Date: Fri, 26 Apr 2002 17:18:06 +0100
- To: "'Www-Zig (E-mail)" <www-zig@w3c.org>
> Could you go into some more detail about this? How would one > define a > query using a URL? In the ARTISTE system all query items are described by RDF schemas, either published by external standards bodies (e.g. the Dublin Core schema at http://dublincore.org/2001/08/14/dces) or written as part of ARTISTE (e.g. the Artiste Core at http://artiste.it-innovation.soton.ac.uk/rdf/ArtisteCore.rdf). Each metadata attribute therefore has a URI e.g. http://dublincore.org/2001/08/14/dces#title Queries are composed using the metadata RDF URIs, and subsequently translated to SQL at each site using a mapping which relates these metadata RDF URIs to individual database table and field names. For example in the collection for Gallery A http://dublincore.org/2001/08/14/dces#title maps to a TitleM field, in the collection for Gallery B the same RDF URI maps a Caption field. I was imaging something similar could be done for SRW if the CQL specification allowed IndexQualifier: [IndexSet.] IndexID | rdfURI A query could then be specified as either dc.creator = "Constable" or http://dublincore.org/2001/08/14/dces#title = "Constable" Another example using the published Bib-1 schema: Contributor = "Montovani, Alberti" or http://www.gils.net/z-tokens.rdfs#21 = "Montovani, Alberti" By the way, is anyone going to be attending the WWW2002 conference at the beginning of May? I wasn't able to attend the recent ZIG meeting but I am going to Waikiki (!) and would appreciate the chance to talk about SRW with anyone who is there. Cheers Alison > > > We have had also had to give thought to specifying queries > and returning > > records which consist of more that simple text - in ARTISTE > every record > > consists of an image and related metadata and queries may consist of > > only an image, only metadata, or a combination of a image > and metadata. > > That, as Mike has just said, is one of the tradeoffs. You > can't do that > using SRW as SRW only works in strings. And until it becomes SSSESRW > (Search, Scan, Sort, Extended Services, Retrieve Webservice) there's a > whole lot more you can't do too. Such as sorting by date or > finding lists > of artists. > > For sophisticated IR, you still need Z39.50 and will continue > to do so > until SRW has put back in the complexity it's currently missing. > > For small projects that have no need of the full scope of > Z39.50, SRW is > (will be) really useful. But your project certainly doesn't > seem to be one > of them. > > UK Academic projects of note are JAFER (Matthew Dovey at Oxford) and > Cheshire (Me, Liverpool in conjunction with UCB) I suspect > that JAFER is > more suited to your needs though as it's Java based, but > Cheshire would > certainly be useful as well. > > Rob > > -- > ,'/:. Rob Sanderson (azaroth@liverpool.ac.uk) > ,'-/::::. http://www.o-r-g.org/~azaroth/ > ,'--/::(@)::. Special Collections and Archives, extension 3142 > ,'---/::::::::::. Twin Cathedrals: telnet: > liverpool.o-r-g.org 7777 > ____/:::::::::::::. WWW: http://liverpool.o-r-g.org:8000/ I L L U M I N A T I
Received on Friday, 26 April 2002 12:18:08 UTC