You can have a data standard and a standard way to exchange data, even 
in the linked data world. I personally don't call it a record. Some 
people might. But I don't think it matters what we call it. And 
application profiles are descriptions of graphs that can include 
constraints that are related to data creation but that do not constrain 
use on the open web. Tom Baker and I will have a discussion paper out 
on that in June, leading up to the discussion at DC2013 [1]. I think 
your knowledge of OWL would be an asset to that discussion, since we 
definitely have not solidified our ideas - we want a lot of input. 
Also, that discussion will of necessity inter-twine with the discussion 
on quality assurance in RDF [2] that is taking place right after the 
DC2103 meeting.


On Sun May 26 10:34:23 2013, Young,Jeff (OR) wrote:
> I'm not saying all installations would have to use the exact same mapping file. They could tweak it to deal with their local idioms.
> I'm not sure what you mean by "future data carrier". Are you talking about some new kinds of closed-world graphs (aka "records")? I also hear people talking about "lightweight abstraction layers". I also hear people talking about "application profiles". I could name others. Same things? A record by any other name would still smell.
> Jeff
> Sent from my iPad
> On May 26, 2013, at 12:56 PM, "Karen Coyle" <[log in to unmask]> wrote:
>> Jeff, I don't think that the question of programming tools is the hard one. There are lots of programming tools. What will be difficult will be understanding and working with the huge variety in data practices - most of which are undocumented, and some of which are idiosyncratic if not down-right unfortunate (such as the addition of ISBNs for different manifestations to a single MARC record, which has been much discussed here). It's one thing to address a data standard - it's another thing to address the actual practice.
>> And as for that practice - I'm not saying that catalogers have been sloppy or remiss. They have worked, often in isolation, to try to satisfy a whole panoply of requirements: requirements related to sharing; requirements related to service to a specific population of users; requirements related to systems functionality that is beyond their control.
>> We need a future data carrier, but I caution that we have to continue to expect a great deal of variability in the data itself. Rather than making "rules" for how to convert from MARC to BIBFRAME, we should be creating flexible conversion routines that libraries and systems can use with knowledge of their own practices and needs, and then we need to talk collectively about coordinating in a shared environment. I actually think that the *intention* of BIBFRAME is to facilitate that flexibility, and maybe we have been remiss in thinking that the LC MARC-to-BIBFRAME process = BIBFRAME. This is why I would like there to be BIBFRAME discussions that are not solely "how to transform MARC." That's only one role that BIBFRAME will fulfill, I hope.
>> kc
>> On Sun May 26 08:33:28 2013, Young,Jeff (OR) wrote:
>>> I suspect that some ILS systems use relational databases. If so,
>>> someone could download D2RQ or one of the more modern R2RML tools and
>>> use that to map their data to the BIBFRAME model. They could then
>>> publish that mapping for other installations of that ILS system to use.
>>> Jeff
>>> Sent from my iPad
>>> On May 26, 2013, at 11:05 AM, "Karen Coyle" <[log in to unmask]
>>> <mailto:[log in to unmask]>> wrote:
>>>> Mac,
>>>> I'm assuming that there will be programs available, just as now there
>>>> are programs for converting from MARC21 to MARCXML. What I am less
>>>> confident about is whether these programs, built for LC's records,
>>>> will give the desired results for other libraries. There is a huge
>>>> difference between a single library with a fairly coherent set of
>>>> practices [*] and the variety of actual data in the range of
>>>> libraries that you work with. What would be great would be the
>>>> establishment of an open source repository where people who make
>>>> modifications or write their own programs could make those programs
>>>> available for others. It may be easier to adapt a program that has
>>>> worked for, say, multi-lingual catalogs than to begin with a program
>>>> designed for a library with only one language of cataloging.
>>>> kc
>>>> [*] I was struck by a remark of Kevin's about the ISBNs that they
>>>> have found the when there are multiples, the first is for the
>>>> hardback. I can very much imagine that being the case in LC's
>>>> catalog, given their purchase and work-flow, but as you know, 1) not
>>>> all systems keep the fields in order 2) not all libraries purchase
>>>> the hardback before the trade paperback. So that's an example of
>>>> something that LC can rely on for their data that may not generalize
>>>> to other libraries.
>>>> On 5/25/13 7:56 PM, J. McRee Elrod wrote:
>>>>> When/if Bibframe is implemented by the national libraries and
>>>>> bibliographic utilities, will there be a publically available Bibframe
>>>>> to NMARC crosswalk for those systems still in MARC?  Would OCLC offer
>>>>> a MARC export?
>>>>> Might prioducing MARC21 records from Bibframe be a niche market for
>>>>> SLC, as is oroducing AACR2 compatible records from RDA, and UKMARC
>>>>> from MARC21?
>>>>> Of what online conversion programs should we be aware?
>>>>> Thanks
>>>>>    __       __   J. McRee (Mac) Elrod ([log in to unmask]
>>>>> <mailto:[log in to unmask]>)
>>>>>   {__  |   /     Special Libraries Cataloguing HTTP://
>>>>>   ___} |__ \__________________________________________________________
>>>> --
>>>> Karen Coyle
>>>> [log in to unmask] <mailto:[log in to unmask]>
>>>> ph: 1-510-540-7596
>>>> m: 1-510-435-8234
>>>> skype: kcoylenet
>> --
>> Karen Coyle
>> [log in to unmask]
>> ph: 1-510-540-7596
>> m: 1-510-435-8234
>> skype: kcoylenet

Karen Coyle
[log in to unmask]
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet