> Amazon are working on providing full text of their books. I would expect
> that eventually they'll want to have it searchable. But your argument is
> also in favour of my position... you likely /don't/ want to return the
> full text of the book, you want to return a chapter or a paragraph or ...
> meaning you want to be able to extract these from the record.

I think the question is in Z39.50, are books going to be stored as
one record or many? If a single book, is there sufficient query
power to do the targeted queries on single smaller units that people
will need?

My feeling is Z39.50 does a good job of querying collections of small-ish
records and does not have a good enough query model to handle large monolithic
(book sized) records with very rich structure. (People may disagree with
this completely.)

*If* the records are generally going to be small-ish, then most concerns
about how to ask for what bits goes away - keep it simple and get the lot.

A single XPath expression is not too bad, but XSLT stylesheets are
problematic as do clients have to supply a full XSLT stylesheet for the
server to process? The client request then becomes large to reduce
a response size. If the XSLT stylesheets are already preloaded, then
you are back to element set names in effect.


But I agree it is certainly possible to use Z39.50 with large records
with rich structure. Z39.50 does have a degree of query power.
Boils down to understanding the user base who would be interested
in Z39.50.

My person feel (having only read the last few mail items and nothing
else on this list for months) is keeping SRW simple is good, or else
potentially suffer the fate of Z39.50 - too complex to implement
and interoperate reliably.


But I don't think my vote counts much as I have not been involved in
CQL for so long that I am really out of all the discussions.

Alan