> Date: Fri, 25 Jul 2003 23:09:55 +0100
> From: Robert Sanderson <[log in to unmask]>
>
> Some questions on Zeerex.  The first two were incontrovertable, but
> the ones below aren't immediately clear from our documentation.

He asks some good questions!

>> 3) The configInfo documentation is unclear if the <default>,
>> <setting> or <supports> tag could be applied to any of the types;
>> the Zing page at
>> http://www.loc.gov/z3950/agency/zing/srw/splain-doc.html lists
>> these as three mutually exclusive sets.  My server will not be able
>> to support caseSensitive=false searching for example.  I'd like to
>> indicate that <setting type="sortCaseSensitive">true</setting> to
>> inform the client that it will not accept case insensitive sorts.
>> I presume its OK to do this?
>
> 3) That sounds sensible!
>    And we should make a note of it.

Seems wrong to me.  Though I admit I don't see a better way, with the
current set of primitives, to express "this server can't do
case-insensitive sorting".  There are at least four possible
approaches: (1) kludge it as he suggests; (2) augment the ZeeRex DTD
to allow an explicit <doesntSupport> statement; (3) specify that in
the absence of a <supports> tag, the client can assume that a facility
is not supported; (4) get him to fix his server so it supports
case-insensitive sorting, then the problem doesn't arise :-)

I can see real drawbacks to all four of these.

>> 4) The documentation is not clear whether <default
>> type="indexSet">...</default> should contain the index schema
>> identifer (http://www.loc.gov/zing/cql/bath-indexes/v1.0/) or the
>> index schema name (bath).
>
> 4) The short form. (as we use the short form everywhere else apart
>    from the attribute that identifies the full URI)

So, Rob, will you fix the document?

>> 5) Does the schemaInfo for the diagnostics record schema need to be
>> specified?  SRW says its mandatory, so it seems redundant to
>> specifiy it.  And I certainly don't want to imply that clients can
>> send SRW searches that request this schema.
>
> 5) You shouldn't ever give the diagnostic schema, as you can never
>    retrieve documents with it.  (Otherwise we run into the problem
>    of real diagnostic vs diagnostic record)

I agree.

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  "When Christ calls a man, he bids him come and die" --
         Dietrich Bonhoeffer.

--
Listen to my wife's new CD of kids' music, _Child's Play_, at
        http://www.pipedreaming.org.uk/childsplay/