Steve, Ros and any others interested,
If you aren't using the Language limit in searches of CURL you can use it to
limit by Code--institution. You just make tag 4 have use=56 instead of
use=54. This works for the java client in mil, haven't tested it for
Cataloging Workstation.
Has anyone got the date limits to work with CURL. After much effort all I
can get is date equals rather than =< and => which is bad enough but index
is for all dates not just publication date. Have tried both use=30 and
use=31 it makes no difference.
In the case of the mil java client:
If you do a author/title search on CURL (use=1000) then the result screen
displays a list of site code followed by some reference number, therefore I
would think the info you want is available in the record returned by CURL
whether it be a brief or a full record. It is really a question of what info
the origin is prepared to present.
It would however be a good item to raise the problems now rather than later
as phase 3 is due to include conformance to bath profile level 0. For
example:
"Complex keyword searches can be expressed using Boolean operators to
connect one or more operands where the operands are constructed using the
searches defined in these specifications."
"Since servers are required to support a minimum number of well-defined
searches for each conformance level, they should be able to process a query
that combines operands that express searches across different indexes (i.e.,
cross index searching)."
The implications of this in terms of redesign of both the cataloging
workstation and the java client are very significant. I would say that if we
want any changes then we have to get them in now.
Cheers Mark
--ooOoo--
Mark Denham, Head of IT Services
Glasgow University Library
Hillhead Street, Glasgow, G12 8QE
Email: [log in to unmask]
Phone: +44 141 330 6765
-----Original Message-----
From: Steve Clews [mailto:[log in to unmask]]
Sent: 31 July 2002 15:58
To: [log in to unmask]
Subject: Re: Marc21 and record downloading.
Hi Ros
That is indeed the issue I was hinting at in my initial email...
I have a file of III emails dated some 2 yrs back, where I pressed helpdesk
on
the inadequacies of the `Origin' display, which as you have spotted cannot
currently pick up the real library source code of the record.
It will currently only tell you that the origin of the record is CURL which
of
course you know already...
Rumour has it the CURL database is considering following RLG practice
which `clusters' their records. That would ameliorate the problem...
eventually.
However a more flexible interface showing library source code of the record
or a limit option to allow selection on source code would be an obvious
improvement...
...this is just the sort of debate that would be useful...
Regards and thanks
Steve
On 31 Jul 2002 at 14:42, Rosalind Pan wrote:
> Here at Durham our database is being converted to MARC21 even as I send
> this, and although I don't have the details to hand, our Collection
> Management staff are less than ecstatic about the single line displays
> etc they are going to be getting in GUICAT Z39.50 searching of CURL
> which is going to be much more cumbersome to use than the old telnet
> route.
>
> I have down as an action point to work through whether the limitations
> are with the INNOPAC implementation of Z39.50 or on the CURL side, so
> any information in this regard would be appreciated as others may
> already have this.
>
> And I would also be interested in software alternatives if any are in
> use.
>
> I wonder if this might be a good topic for the User Group, as an
> optional session?
>
>
> Ros Pan
> Durham.
>
> --
> # University of Durham Name: Ros Pan #
> # Main Library Occupation: Sub Librarian #
> # Stockton Road Voice: (0191) 374 3411 #
> # Durham Fax. No.: (0191) 374 7481 #
> # DH1 3LY E-Mail: [log in to unmask] #
--
Steve Clews
Sub-Librarian, Information Strategy
University of London Library [log in to unmask]
Senate House, Malet St Phone: 020 7862 8452
London WC1E 7HU http://www.ull.ac.uk
|