Date:         Tue, 18 Jun 2002 00:04:24 +0200
Reply-To:     "Z39.50 Next-Generation Initiative" <[log in to unmask]>
Sender:       "Z39.50 Next-Generation Initiative" <[log in to unmask]>
From:         Sebastian Hammer <[log in to unmask]>
Subject:      Re: Betr.: Re: result set model for srw
Comments: To: "Z39.50 Next-Generation Initiative" <[log in to unmask]>,
          [log in to unmask]
In-Reply-To:  <[log in to unmask]
              .ac.uk>
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 22:34 17-06-2002 +0100, Robert Sanderson wrote:
> > You mean like the XML format Eliot and I came up with for Type-101 queries
> > (see http://www.gils.net/search.html)?
>
>Like that yes. But not so complex as we don't need <relation> and
><semantic> and <NameOfAttributeIJustInventedOffTheTopOfMyHead> etc.
>
>Take current CQL and wrap appropriate XML tags around the various bits in
>it.

Some benefits of Rob's proposal:

o Syntax is unambiguous and easy to verify even for non-CS people (I don't
think that tools for verifying expressions against BNF specs are commonplace).

o It's easy to parse.

o It's easy to construct in software.

o It can be manipulated in XSLT without resorting to gymnastics.

o It's neatly extensible in the future without kludging stuff into a human
syntax.

o People will never make the mistake of allowing their users to type it in
(and hence give Matthew heart palpitations).

Note btw that XQuery (or did, last I looked) postulates both a
human-friendly syntax and an XML encoding.

--Sebastian
--
Sebastian Hammer, Index Data <http://www.indexdata.dk/>
Ph: +45 3341 0100, Fax: +45 3341 0101