Date:         Fri, 23 May 2003 18:23:29 +0200
Reply-To:     "Z39.50 Next-Generation Initiative" <[log in to unmask]>
Sender:       "Z39.50 Next-Generation Initiative" <[log in to unmask]>
From:         Theo van Veen <[log in to unmask]>
Subject:      Re: Betr.: Re: Metaresponse
Comments: To: [log in to unmask]
Content-Type: text/plain; charset=US-ASCII

With respect to record I understand the confusion. There are two
different situations:

1) There may be a SRU record identification that can be used to
retrieve a SRU record conform the SRU protocol (like resultset and
recordPosition)
2) There may be within the record an internal recordid or link to the
record in the native interface - such a field is not meant for use by
SRU but is completely the responsibility of the client and server
application.

Like the SRU part of the application should - in my opnion - not bother
about what is inside the actual record, the same could apply to the
query: the SRU application hands the query over to the  search engine
without bothering what is inside.

I did not change opinion in those 5 minutes but there were at the same
time to discussions going on on two different kind of recordid's.

Theo

>>> [log in to unmask] 05/23 3:44 nm >>>
> >B) Why not just use the existing mechanism of CQL?
> >...&query=identifier%="foo1"
> This could also be an option, although I personally prefer to
> distinghuish between processing SRU/SRW parameters (as part of the
> SRU-protocol) and processing the contents of the metadata record (as
> part of the actual applications). This is also the reason why I
never
> like resultsetid to be part of the query instead of being a separate
> parameter.

5 minutes ago you were arguing in favour of having the metadata be part
of
the record. But you don't want to use normal CQL to search it and
would
rather process metadata separately?  Doesn't this strike you as odd?

> >> valuable. The concept of DCX is explained below and I will submit
a
> >> proposal to the DCMI to adopt this concept. I would like to
propose
> >Changing the requirement to support DCX over DC is a little too far
> >though, as it could require a LOT of work on the client side to
> > handle
> I would expect the opposite, because you only have to neglect terms
you
> do not know. When using stylesheets you do not have to do anything
to
> neglect the terms you do not know.

Err, there are 15 fields in DC that have no nesting. There are
potentially
infinite nested fields in DCX.

What happens if you're /not using stylesheets/?  Last time I looked at
the
specs, it didn't say 'You must implement this using XSL' which is
effectively what you're saying by making DCX mandatory. Which cuts out
any
environment that doesn't have an XSLT engine.  I'm not prepared to do
that.

> >sensibly.  DC as mandatory is there to enable a minimum standard,
not
> >necessarily one that can be used for all purposes.
> In this respect DC and DCX do not differ.

Yes they do.  DCX could not be described as 'minimum'.
If the DCMI agree to support it, then I will too, but not as
mandatory.

Rob

--
      ,'/:.          Rob Sanderson ([log in to unmask])
    ,'-/::::.        http://www.o-r-g.org/~azaroth/
  ,'--/::(@)::.      Special Collections and Archives, extension 3142
,'---/::::::::::.    Twin Cathedrals:  telnet: liverpool.o-r-g.org
7777
____/:::::::::::::.              WWW:  http://liverpool.o-r-g.org:8000/

I L L U M I N A T I