RE: [search-ws] queryn: Some further comments

["LeVan,Ralph" <levan@oclc.org>]

I believe you've correctly summarized my sense of the conversation.  I'd
prefer that cql-form not be in the standard, but I do believe we need to
document the queryType definition requirements and in the standard we
should point to a web page where we will document the know queryTypes.

If we (Tony) wants to support Booleans before or after, then I think we
should have a separate queryType for each syntax.  But, once he's gone
through the effort of describing how to create forms for one of those
syntaxes, I suspect he'll not see a need to do the same for the other
syntax.  Why would a form creator care?  As long as the rule is clear
about where the Boolean goes, I suspect the forms will get created
correctly.  Make the documentation unambiguous about where the Boolean
goes in relation to the terms and I think the problem is solved.

Ralph

> -----Original Message-----
> From: Matthew Dovey [mailto:m.dovey@jisc.ac.uk]
> Sent: Monday, December 20, 2010 3:15 PM
> To: Ray Denenberg, Library of Congress; LeVan,Ralph; Hammond,Tony;
> 'OASIS SWS TC'
> Subject: RE: [search-ws] queryn: Some further comments
> 
> > "The queryN parameter is peculiar to the cql-form queryType"
> 
> Can we go back a few steps, and check what we agree on.
> 
> Where I think we are at is:
> 
> i) we agree that we change the spec to allow the value queryType to
determine
> which parameter(s) contain the query. The default is that the query
will be
> passed via a parameter called "query", but a specific queryType can
specify
> otherwise.
> 
> ii) we agree that we should introduce a new queryType to handle Tony's
use
> case. This queryType will use different parameters to pass the query
than the
> "query" parameter.
> 
> Where we still need to agree:
> 
> i) the name for this new queryType - or are we happy with "cql-form"
> 
> ii) the status of this queryType - normative, non-normative but
included in spec;
> Nature specific extension etc. I would favour that this is at least a
non-normative
> part of the spec.
> 
> ii) the syntax for this queryType
>   a) the need for "queryn" parameter - although I think we've probably
reached
> the conclusion it is needed
>   b) leading versus trailing Booleans
>   c) exact names for the parameters (although q[i|r|t|b]<n> seems the
current
> candidate).
> 
> 
> Matthew
> 
> 
> 
> 
>