W3C home > Mailing lists > Public > public-lld@w3.org > March 2011

RE: Question about MARCXML to Models transformation

From: Young,Jeff (OR) <jyoung@oclc.org>
Date: Wed, 9 Mar 2011 15:09:38 -0500
Message-ID: <52E301F960B30049ADEFBCCF1CCAEF590BBB784A@OAEXCH4SERVER.oa.oclc.org>
To: "Diane I. Hillmann" <metadata.maven@gmail.com>
Cc: "Ross Singer" <ross.singer@talis.com>, <public-lld@w3.org>
Diane,

I downloaded the code and think I found the core MARC to XC FRBR logic
in the xc.mst.services.transformation.TransformationService class. It's
a big class with unclear dependencies, but looks useful as a source of
clues. It's not readily apparent that most MARC records don't contain
many-to-many relationships, though, so thanks for the clue. 

Jeff

> -----Original Message-----
> From: Diane I. Hillmann [mailto:metadata.maven@gmail.com]
> Sent: Wednesday, March 09, 2011 12:52 PM
> To: Young,Jeff (OR)
> Cc: Ross Singer; public-lld@w3.org
> Subject: Re: Question about MARCXML to Models transformation
> 
> 
>   Jeff,
> 
> Most MARC records don't contain many-to-many relationships, except for
> an aggregated container (festschrift, some music records, multiple
> version records, etc.).  Those are rather easily flagged.  You should
> try out the MST, and take a look, rather than listen to me go on about
> it! (http://www.extensiblecatalog.org/).
> 
> Diane
> 
> On 3/9/11 12:28 PM, Young,Jeff (OR) wrote:
> > Diane,
> >
> > There is a many-to-many relationship between E&M. Does the toolkit
> try
> > to detect the presence multiple expressions in a MARC record or does
> is
> > assume there is only one?
> >
> > Jeff
> >
> >> -----Original Message-----
> >> From: public-lld-request@w3.org [mailto:public-lld-request@w3.org]
> On
> >> Behalf Of Diane I. Hillmann
> >> Sent: Wednesday, March 09, 2011 12:00 PM
> >> To: Ross Singer
> >> Cc: public-lld@w3.org
> >> Subject: Re: Question about MARCXML to Models transformation
> >>
> >>
> >>    It might be good to note here that the strategy used by the
> >> eXtensible
> >> Catalog Metadata Services Toolkit is to parse the MARC records out
> > into
> >> a WEMI structure, then go back and determine which Works are the
> same
> >> (and from there, Expressions, etc.), rather than try to determine
> that
> >> as part of the initial process.  This seems to me a good strategy.
> >>
> >> Diane
> >>
> >> On 3/9/11 10:53 AM, Ross Singer wrote:
> >>> On Wed, Mar 9, 2011 at 10:27 AM, Richard Light
> >>> <richard@light.demon.co.uk<mailto:richard@light.demon.co.uk>>
> > wrote:
> >>>
> >>>      This is probably my bibliographic ignorance coming out, so
> >>>      apologies if that's the case.  I was assuming that "different
> >>>      URLs" = "different Expressions". However, from the discussion
> > I'm
> >>>      realising that the issue is about how to tease out WEMI from
> >>>      records which lack this world view.
> >>>
> >>> Richard, it's understandable - it's a mess!
> >>>
> >>> But this is sort of where I'm getting at.  Our best laid plans
> still
> >>> need to account for the 214 million records (just using Worldcat
as
> >> an
> >>> example, meaning there are considerably more) that already exist.
> >>>
> >>> I think the day when we hit the 24 hour period that more RDA
> >> resources
> >>> were created than AACR2 we can begin to start thinking about how
> >>> RDA/FRBRer/etc. begins to shape the way we create linked data.
> > Until
> >>> then, all we have to work with is what we've got.
> >>>
> >>> -Ross.
> >>
> >
> >
> 
Received on Wednesday, 9 March 2011 20:10:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 March 2011 20:10:32 GMT