Print

Print


On 7/22/14 9:59 PM, Stuart Yeates wrote:
> I have a healthy distrust of such lists, for two reasons.
>
> Firstly they never seem to include local things I care about, such as
> http://mshupoko.natlib.govt.nz/mshupoko/index.htm (Interestingly,
> subject headings like that become much more interesting in the linked
> data world, but that's another story.)
>
> Secondly because in a linked data world, abbreviated codes that stand
> for something we will already have URIs for is a redundant source of
> confusion. By all means use the abbreviated codes for human-readable
> mnemonics, but use the full URI for data matching and semantics and make
> it clear that it's the URI not the abbreviated code that carries meaning.

In the list given there is an entry for "msc" which is the Mathematics 
Subject Classification of AMS / Zentralblatt Math. There are actually 
three versions which are not compatible:

MSC2010 - http://www.ams.org/msc/msc2010.html
MSC2000 - http://www.ams.org/msc/msc.html
MSC1991 - http://msc2010.org/msc1991/

See, for example, 2000->2010 mappings at 
http://www.ams.org/msc/conv.html. Unfortunately AMS and Zbl don't given 
nice URIs for theses schemes but it does seem clear that the string 
"msc" is not adequate.

Rgarding locally created URIs for schemes that aren't in a central 
registry and don't have an accepted URI from source: It is much easier 
to migrate data from two sources using cornell:MSC2010Classification and 
harvard:MSC2010Classification to standard:MSC2010Classification, than it 
is to untangle incompatible data in, say, MSC2010 and MSC1991 both 
listed as standard:msc. So it seems better to err on the side of 
encouraging local classes to be used when in doubt or when there is not 
yet community agreement on a specific class URI.

Cheers,
Simeon


> cheers
> stuart
>
> On 07/23/2014 09:53 AM, Ford, Kevin wrote:
>> I just wanted to chime in two things here:
>>
>> 1) I want to acknowledge my agreement with the below but also add that
>> this is one of those areas that I believe requires a greater education
>> effort (on all of our parts).  We need to communicate how other
>> communities/implementers can do this responsibly; and
>> 2) A hybrid solution is not only possible but a reality if you
>> consider contemporary MARC practice.
>>
>> More about two: We already maintain a healthy list of classification
>> "sources:" http://www.loc.gov/standards/sourcelist/classification.html
>>
>> The codes are meant to be used in MARC Bib 084 $2.
>>
>> Although we never formally published it, we actually turned that list
>> into RDF and added it to ID.LOC.GOV.  We didn't bring it to completion
>> because we noted there were two distinct ways to publish the list with
>> respect to possible use and never quite resolved that question before
>> our attentions were turned to bigger things.  Regardless, the point
>> being is this: we could actually leverage a fair amount of existing
>> work to create a solid base set of URIs to unambiguously identify the
>> "source" of the classification.  So not just a few big ones could be
>> immediately covered, but many smaller ones too (there are at least 149
>> entries on that list).
>>
>> Yours,
>> Kevin
>>
>> p.s.  Nate, before someone else points this out :) , and do know I
>> actually wasn't looking for it (it was news to me), but the Chinese
>> library classification (code: clc) is actually on the list!
>>
>>
>>
>>> -----Original Message-----
>>> From: Bibliographic Framework Transition Initiative Forum
>>> [mailto:[log in to unmask]] On Behalf Of [log in to unmask]
>>> Sent: Tuesday, July 22, 2014 5:13 PM
>>> To: [log in to unmask]
>>> Subject: Re: [BIBFRAME] Deproliferation of Predicates
>>>
>>> Although the question was addressed to Rob Sanderson, I'm going to take
>>> the liberty of opining:
>>>
>>> Let's invent a local scheme:
>>>
>>> local:ScientificDataSetClassification rdfs:subClassOf
>>> bf:Classification .
>>>
>>> local:EnvironmentalScienceDataSet a skos:Concept ;
>>>     skos:inScheme local:ScientificDataSetClassification .
>>>
>>> local:PhysicsDataSet a skos:Concept ;
>>>     skos:inScheme local:ScientificDataSetClassification .
>>>
>>> and so forth. With that published, we can use it in Linked Data.
>>> Kevin Ford
>>> made the point in another thread that it's often a good practice to
>>> use local
>>> URIs for things about which you want to make assertions, but which you
>>> don't "own" yourself. Here's an example of why. If Bibframe were to
>>> adopt
>>> and support a new bf:DataSetClassification corresponding to the
>>> above, we
>>> could switch over to using it, and we could republish the RDF with
>>> the first
>>> statement replaced:
>>>
>>> local:ScientificDataSetClassification owl:sameAs
>>> bf:DataSetClassification .
>>>
>>> but we wouldn't have to completely rework all of our metadata to respond
>>> to a shift in Bibframe. From my perspective, the move here is to
>>> shift the
>>> evolution of such entities (classification schemes and identifiers
>>> are good
>>> examples) out of our metadata and into the community that supports it,
>>> while at the same time reaping the benefits that come with having
>>> identifiers
>>> and not labels (literals) as values in our metadata. When anyone can
>>> publish a
>>> new classification scheme (for example) without disrupting other people
>>> using Bibframe, space is opened up for that kind of evolution in a
>>> healthy
>>> way, without losing the benefits of a controlled process for the
>>> evolution of
>>> core ideas (which, in this example, would be the generic notion of a
>>> classification scheme). (For example, if I worked at a small public
>>> library that
>>> couldn't care less about scientific data sets, I'm not called upon at
>>> any point in
>>> the above-imagined process to do anything to my metadata.) At the same
>>> time, the shared model grows no more than is really needed.
>>>
>>> ---
>>> A. Soroka
>>> The University of Virginia Library
>>>
>>> On Jul 22, 2014, at 4:47 PM, "Trail, Nate" <[log in to unmask]> wrote:
>>>
>>>> Rob,
>>>>
>>>> Well, we seem to have a list of a handful of the big classification
>>>> schemes
>>> (in the current bibliocentric world), but nothing from China or from
>>> some
>>> new data format or wherever else; what do we do with their schemes when
>>> we express their data in our systems, before they become so generally
>>> accepted in BIBFRAME that they get their own? How do we express a local
>>> classification scheme?
>>>>
>>>> Nate
>>>>
>>>> From: Bibliographic Framework Transition Initiative Forum
>>> [mailto:[log in to unmask]] On Behalf Of Robert Sanderson
>>>> Sent: Tuesday, July 22, 2014 4:33 PM
>>>> To: [log in to unmask]
>>>> Subject: Re: [BIBFRAME] Deproliferation of Predicates
>>>>
>>>>
>>>> Hi Nate,
>>>>
>>>> -------------------------------------
>>>> However, even better given the query optimization scenario would be:
>>>>      _:x bf:classification [ a bf:NlmClassification ; value "123" ]
>>>> -------------------------------------
>>>> Haven't you gone here from deproliferating properties to proliferating
>>> classes?  If we did this, wouldn't we also have to have these classes:
>>>>
>>>> I have, yes :) But at the same time, getting rid of the string
>>>> literals in
>>> scheme, and indeed, the scheme property all together.  So a net
>>> reduction in
>>> the model, and a net increase in simplicity and ease of querying.
>>>>
>>>> and you'd still probably need the generic bf:Classification and a
>>>> way to say
>>> what scheme it represents, as we develop or accept new mechanisms to
>>> organize new types of material that don't rise yet to the level of
>>> having their
>>> own class?
>>>>
>>>> I don't follow that, I'm afraid. Why wouldn't we (the community)
>>>> give new
>>> Classification types their own subclass?
>>>>
>>>> Rob
>>>>
>>