Dear Phil,

Great to hear that the code worked for you.

> One question, how would you like
> comments/questions sent in?  
-- Asking this question makes me fear the answer.  Post to the listserv.  If it's a bug, we can dispense with it quickly, but it might be a bigger issue for discussion.

> the 700s for additional authors are being attached to the Instance not
> the Work.  
-- This is actually by design.  In part this is because we've not divied up the relators codes and associated them with BIBFRAME Works or Instances, thereby allowing us to determine with which resource (the Work or Instance) the name is associated.  For example, I would think an Actor or a Puppeteer be associated with an Instance versus a Work.  And we've not divied up the relators, in part, because we (LC) do not have a lot of relators codes in our data (a hair under 800K, to be precise) so it has lacked some immediacy (in these early dev days).  And, again because we do not have a lot of relators codes in our data, it becomes even more difficult to determine with which resource (Work or Instance) the name is related.  Look at a music or moving image resource and suddenly all those 700s and 710s being associated with BIBFRAME Instance resources make perfect sense. 

Now, if you have records that have employed relators codes, rest assured that any names with relators codes like "cre" or "aut" will be associated with the Work. Is that the case?

Regardless, your overall point is taken.  Not all 700s and 710s should be associated with BIBFRAME Instance resources, and should instead be with the Work.  *That* is going to take some figuring out.

Warmly,

Kevin




> -----Original Message-----
> From: Philip Schreur [mailto:[log in to unmask]]
> Sent: Friday, December 07, 2012 4:55 PM
> To: Bibliographic Framework Transition Initiative Forum
> Cc: Ford, Kevin
> Subject: Re: [BIBFRAME] Early Experimentation Code Available
> 
> Kevin,
> 
> Thanks so much for making this available!  Since we've been cataloging
> in RDA from when testing began, it'll be fun to take all of the records
> and do a conversion.  One question, how would you like
> comments/questions sent in?  For instance (we used the XQuery version),
> the 700s for additional authors are being attached to the Instance not
> the Work.  In our first example, they were compilers and I thought
> there may have been some confusion there but in the second they were
> just straightforward authors.
> 
> Phil
> 
> On 12/7/2012 7:04 AM, Ford, Kevin wrote:
> > We're making code available that will permit programmers and
> developers to better understand how MARC Bibliographic records can
> transform to BIBFRAME resources.  The code is available at:
> >
> > https://github.com/lcnetdev/marc2bibframe
> >
> > There are two versions: one in Python and one in XQuery.  The Python
> version produces JSON files for viewing in a Simile Exhibit
> presentation (also included).  The XQuery version outputs RDF (as
> RDF/XML, N-triples, or JSON).  Although the Python code expects to be
> invoked from the command line, the XQuery code can be invoked using
> Oxygen XML or the Eclipse IDE, in addition to a few other methods.
> >
> > The XQuery is the product of Network Development and MARC Standards
> staff at LC; the Python version is the work of Zepheira.  The two
> transformations were developed independently of each other.  They
> therefore do not "split" MARC Bibliographic records into BIBFRAME Works,
> Instances, Authorities, and Annotations in the exact same way.
> >
> > Both are subject to change.  Neither is canonical.  They are very
> much works in progress.  As such, the names of properties and
> classes/entities are in flux and will invariably change in many cases.
> Sometimes, developers just need to do something to keep going, even if
> it is subject to alteration later (and often is modified).
> Nevertheless, we want to make these available for evaluation and to
> stimulate further conversation.
> >
> > We are presently working on a way to better expose the output of
> these transformations to a wider audience.  In the next several weeks,
> we hope to provide a service that will permit individuals to submit
> their own batch of MARC Bibliographic records for transformation.  This
> way, you can see what *your* data might look like conforming to the
> BIBFRAME model.
> >
> >
> > --
> > Kevin Ford
> > Network Development and MARC Standards Office Library of Congress
> > Washington, DC
> 
> 
> --
> Philip E. Schreur
> Head, Metadata Department
> Stanford University
> 650-723-2454
> 650-725-1120 (fax)