Re: [search-ws] Another Kind of Facet?

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

How about something like this:

 <facetedResults>
     <dataSource>  <!-- first data source
             <dataSourceLabel>  <!-- description of the datasource -->
             <baseURL> http:// .......   <!-- URL of the datasource -->
           <facets>
                 <facet>  <!-- first facet for this datasource
                           <displayLabel>  <!-- label for this facet -->
                           <index>
                          <relation>
                                <!-- index and relation optional. Query may 
be suppled instead for each term as seen below -->
                          <terms>
                                  <term>   <!-- first term for this facet
                                         <actualTerm>
                                         <query>     <!-- (optional) query 
for this facet/term -->
                                </term>
                                  <term> <!-- next  term for this facet
                                         <actualTerm>
                                         <query>     <!-- (optional) query 
for this facet/term -->
                                </term>
    ........
                         </terms>
              </facet>
              <facet> <!-- next facet for this datasource
                      .....................
             </facet>
        <facets>
    </dataSource>
    <datasource>  <!-- next data source
 ......
   </datasource>
</facetedResults>




----- Original Message ----- 
From: "LeVan,Ralph" <levan@oclc.org>
To: "Ray Denenberg, Library of Congress" <rden@loc.gov>; 
<search-ws@lists.oasis-open.org>
Sent: Thursday, March 19, 2009 9:02 AM
Subject: RE: [search-ws] Another Kind of Facet?


> That part I understood (I think).  My suggestion was
> that these different databases are accessed via
> different base URLs, so if you supply a base URL for
> each, along with facet information specific to that
> database, then they can construct the query.  I'm
> not sure what I'm missing.

The same query might not have been used to search each component
database, so the client might not know how to construct the query.  For
instance, the client might have specified the serverChoice index and the
federating database technology might then have chosen a different index
to search a component.  Simply returning the base URL to the component
will not tell the client how to formulate the search of the component.
The client must be given the complete URL for the search.


> Your suggestion to supply a human-friendly
> ("non-interoperable") name and also supply the query
> which would access the facet associated with that
> name is exactly what I proposed originally and it
> did not go over well in the SRU discussion list.
>
> Maybe you'd like to revive that line of discussion
> on the SRU list?

When we get closer to a proposal, I'll do that.  I think we have a good
use case here.  I am implementing this logic now (literally), so we
can't be accused of fictitious examples.

Ralph


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