RE: Betr.: RE: Bib-2 and the DC-Lib

Barbara,

You are right, most relations are defined. 
My concerns come from the fact that I have a quite different kind of application in mind than the conventional way of searching. Most people want to do well-defined searches in well-described databases with well-defined attributesets  etc. I have an application in mind where the list targets is itself the result of a search and I want do to a distributed search in that automatically generated list of (unknown) targets. I have made such an application for the SRU-protocol (you can find a demo of it via the ZiNG-page) but I am afraid it is getting harder and harder to keep this concept supported.

Theo


>>> Shuh Barbara <barbara.shuh@nlc-bnc.ca> 23-04-02 22:09 >>>
Theo

You said, in closing, in one of your postings this morning:
"It would be very convenient if there is also a well defined relation
between the access points and originating fields expressed in DC-Lib or
MARC."

I don't know if you have looked at Bib-2, but all the way through the
document, there are references to the MARC 21 and the UNIMARC tags to which
the Bib-2 access points relate.

As well, most Bib-2 access points are based on the Z39.50 Cross-Domain
Attribute Set access points, which, at the time that attribute set was
developed, mapped to Dublin Core elements.  The problem is that Dublin Core
has evolved since then, so the fit isn't as good...


Barb
______________________
Barbara Shuh
Library Network Specialist
National Library of Canada
Phone:  (613) 995-1701  Fax:  (613) 943-1939
E-mail:  barbara.shuh@nlc-bnc.ca 


-----Original Message-----
From: Theo van Veen [mailto:Theo.vanVeen@kb.nl] 
Sent: Tuesday, April 23, 2002 9:52 AM
To: azaroth@liverpool.ac.uk; Shuh Barbara
Cc: Robina.Clayphan@bl.uk; www-zig@w3.org 
Subject: RE: Betr.: RE: Bib-2 and the DC-Lib


I think we agree, but I see things from the perspective of SRU/SRW, CQL and
DC-Lib. Iexplain it below and give a suggestion.

We have metadata elements in DC-Lib/XML and MARC other formats. Those
metadata elements do represent something like title or author etc.
We can index those metadata, the access points are related to their origine
like dc.title in xml or tag 245 in MARC. There does not have to be a one to
one relationship between the original metadata fields and the access points
as different metadata elements can be combined into single access points.

When doing a search in an index there must be a translation from the index
type that the user asks for and the access point that is defined in that
index. For Z39.50 we have the attribute set that relates access points to
index fields. In CQL we have not and it has been suggested by Ralph Levan to
use things like DC.title and DC.author in a DC attribute set and use
something like Bib1.title for a Bib1 attribute set.

This starts to become very confusing I think. The way we define application
profiles to define metadata elements that describe objects, like we do in
DC-Lib, could be used also to define an application profile for  index
fields (lets say DC-Bib) . And we need to define in such a profile the exact
mappings to Bib-1 and Bib-2. It would be very convenient if there is also a
well defined relation between the access points and originating fields
expressed in DC-Lib or MARC. Regardless of the abstraction level  we may
expect that a title search is related to the title in the original records.

Theo


>>> Shuh Barbara <barbara.shuh@nlc-bnc.ca> 22-04-02 18:06 >>>
Rob's right.

The modular design of the Z39.50 attribute architecture allows for the use
of multiple attribute sets in one query.  If there are aspects of DC-Lib
profiled data that go beyond that which is carried in MARC-like databases,
then there could be a DC-Lib Attribute Set, if there is a body of interested
persons willing to go through the exercise of developing such a set... (You
may note that, at the present time, I don't see myself as part of that
group;-)

Barb
______________________
Barbara Shuh
Library Network Specialist
National Library of Canada
Phone:  (613) 995-1701  Fax:  (613) 943-1939
E-mail:  barbara.shuh@nlc-bnc.ca 


-----Original Message-----
From: Robert Sanderson [mailto:azaroth@liverpool.ac.uk] 
Sent: Monday, April 22, 2002 10:33 AM
To: Theo van Veen
Cc: Kevin.Gladwell@bl.uk; Shuh Barbara; www-zig@w3.org 
Subject: Re: Betr.: RE: Bib-2 and the DC-Lib


> The interoperability problem is that: "when the DC-lib community find
> searches which cannot be created in BIB2 they have to make their own
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> attribute set, DC-Lib-1 or whatever, which allows for these extra
> searches."

How is that an interoperability problem?

There isn't two ways to do a search, there's the BIB2 way for all that 
BIB2 defines and the extras are defined in DC-Lib-1.

There is no overlap, therefore there is no interop problem?

Rob

-- 
      ,'/:.          Rob Sanderson (azaroth@liverpool.ac.uk)
    ,'-/::::.        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

Received on Wednesday, 24 April 2002 05:48:06 UTC