Re: goals wrt FRBR/RDA (was Re: Works and instances)

Karen,

On 08/01/2013 23:18, "Karen Coyle" <kcoyle@kcoyle.net> wrote:

> I guess I'm the "anti-Gordon" because I am a confirmed FRBR doubter.
That conjures up an image ;-)

> It seems to me that schema.org needs to mark up bibliographic data that is
> exposed on the web.
Totally agree, however we have to be realistic that WEMI has heavily
conditioned the language of discussion in our domain.

What I was implying with my ³If we cannot demonstrate how you would publish
your data using [extended] Schema.org, that is currently in
FRBR/RDA/Marc/ISBD/etc., we will not have completed our mission.² statement
was that we need to satisfactorily advise users, of our proposed Schema.org
extensions, how they would describe entities and their relationships that
they currently are describing using FRBR WEMI terms.  Not forgetting to
emphasise that Schema is a vocabulary for sharing data with the wider web
that should complement any other metadata format, not replace it.
 
> Remember that schema.org marks up *displays* -- it
> is not designed to mark up data as it is stored in databases.
Whilst I agree that the initial focus of Schema is embedding structured data
mark up within html, the fact that it is a vocabulary for Œstructured data¹
means that its benefit can spread far beyond the web page.  The search
engines are harvesting the data into their internal datasets ­ Google are
using it as one of the sources to populate their Knowledge Graph.  Schema
are evolving towards using the vocabulary in RDFa, as much as microdata.
RDFa naturally is available as RDF which means that it is ready for hosting
in data stores or making available as dumps ­ as OCLC have done for the most
highly held WorldCat resources.

> If we get to a point where there are displays that use the FRBR entities
> explicitly, then we should add those to schema.org.
In the mixed vocabularies world of Linked Data it would be perfectly
possible to mix [course grained] Schema descriptions with [more fine
grained] FRBR/RDA/BIBFRAME data in the same page/dataset.

> At the moment the only usages I am aware of (other than VTLS cataloger view,
> which is not
> exposed publicly) is the presentation of works in Open Library and
> LibraryThing (and maybe goodreads?) So we could look to those for fields
> that are not yet covered by book, such as "work title". I caution
> against getting ahead of ourselves, however, since we do not at this
> stage know beyond this what "FRBR" displays are since actual FRBR data
> is not being displayed. (And, if BIBFRAME goes forward as planned, may
> be less likely to be built into library bibliographic work.)
Work oriented datasets have been the ambition of many for a long time and
there is, and has been, much effort put into the algorithms for clustering,
grouping  and other methods for inferring a Work from the manifestation
oriented data that has been produced by the library community.

I believe that these efforts, plus examples such as provisional results from
BIBFRAME, when added to the need to address the Work oriented view of the
average public user of search engines; will result in progress on this front
sooner than you imagine.  You cite the fact that there are no Works in
WorldCat in support of argument to not focus too much on Works at this
stage.   It is my personal hope that that situation will not continue for
much longer.

So I agree that we should look to the few examples of the presentation of
Works, and at least provide the properties in our recommendations to enable
the description of a relationships between CreativeWorks that could be
loosely described as Work/manifestation relationships.
> 
> Perhaps we need to gather examples of displays that could be coded with
> schema.org. 
This is an excellent way to test Schema.org, and our extension proposals.
Exactly the way OCLC identified some of the issues,  that identified to me
the need for this group in the first place, by trying to mark up a WorldCat
resources.  As I rehearsed in my blog post on the topic a while back
<http://dataliberate.com/2012/11/the-correct-end-of-your-telescope-viewing-s
chema-org-adoption/>, we need to start by trying to describe our resources
using schema to identify the gaps in the vocabulary ­ not to step through
every class and property in our favourite ontology and try to replicate it
in Schema.

~Richard.
> 
> OCLC gives us one set. I could Another set could come from
> the labeled display like LC (from old catalog):
> 
> LC control no.:         96002325
> LCCN permalink:         http://lccn.loc.gov/96002325
> Type of material:       Book (Print, Microform, Electronic, etc.)
> Personal name:  Twain, Mark, 1835-1910.
> Uniform title:  Novels. Selections
> Main title:     The complete Tom Sawyer / Mark Twain.
> Published/Created:      New York : Gramercy Books, 1996.
> Description:    ix, 294 p. : ill. ; 23 cm.
> ISBN:   0517150786
> Contents:       The adventures of Tom Sawyer -- Tom Sawyer abroad -- Tom
> Sawyer, detective.
> Subjects:       Sawyer, Tom (Fictitious character) --Fiction.
>         Finn, Huckleberry (Fictitious character) --Fiction.
>         Adventure stories, American.
>         Humorous stories, American.
> LC classification:      PS1302 1996b
> Dewey class no.:        813/.4
> 
> As well as Amazon (which you can look up), and including a variety of
> citation formats:
> 
> Twain, Mark, John C. Gerber, and Paul Baender. 1982. The adventures of
> Tom Sawyer. Berkeley: University of California Press.
> 
> In fact, it now occurs to me that the examples that I gave of
> commonEndeavor may need to be re-thought in terms of display. Here's a
> display example from the schema.org Book documentation page:
> 
> <span itemprop="name">The Catcher in the Rye</span> -
>   <link itemprop="bookFormat" href="http://schema.org/Paperback">Mass
> Market Paperback
> by <a itemprop="author" href="/author/jd_salinger.html">J.D. Salinger</a>
> 
> kc
> 
> 
> On 1/8/13 2:21 AM, Gordon Dunsire wrote:
>> > Jodi
>> >
>> > Comments inline Š
>> >
>> > Cheers
>> >
>> > Gordon
>> >
>> > *From:*Jodi Schneider [mailto:jodi.schneider@deri.org]
>> > *Sent:* 07 January 2013 17:25
>> > *To:* Gordon Dunsire
>> > *Cc:* public-schemabibex@w3.org
>> > *Subject:* goals wrt FRBR/RDA (was Re: Works and instances)
>> >
>> > On 7 Jan 2013, at 08:08, Gordon Dunsire wrote:
>> >
>> >     Does this provide a way forward forschema.org <http://schema.org>to
>> >     interoperate with FRBR/RDA?
>> >
>> > Perhaps I'm being a bit pedantic, but I don't think I'd state this as a
>> > goal -- at least not directly.
>> >
>> > GD: As Richard said ³If we cannot demonstrate how you would publish your
>> > data using [extended] Schema.org, that is currently in
>> > FRBR/RDA/Marc/ISBD/etc., we will not have completed our mission.²
>> >
>> > Rather, I'd ask:
>> >
>> > - What qualities of library and bibliographic data are needed for search
>> > engines to serve endusers?
>> >
>> > - How does FRBR/RDA promote these qualities? What qualities of the data
>> > needs to be preserved?
>> >
>> > GD: By ³qualities², do you mean attributes and relationships? I think we
>> > have to assume that FRBR/RDA promotes those attributes and relationships
>> > (i.e. provide a semantic model) that serve endusers, given that FRBR is
>> > focussed on user tasks, and we should give credit to the international
>> > processes that determined those attributes and relationships. So I would
>> > say that we know what qualities serve endusers in bibliographic resource
>> > discovery systems; the question is to what extent search engines support
>> > them and, most importantly, to what extent ³amateur² cataloguers (and,
>> > indeed, inference engines) can supply the bibliographic data.
>> >
>> > To me, going back to use cases would really help. If anybody has time to
>> > translate the problems/solutions into specifics, with our use cases, I
>> > think that would bring us forward.
>> >
>> > We currently have a number of use cases here:
>> >
>> > http://www.w3.org/community/schemabibex/wiki/Use_Cases
>> >
>> > Among them are to describe a FRBR Item and a FRBR Manifestation:
>> >
>> > ·1.4Use case: Describe a FRBR Item (e.g. a single identifiable book)
>> > 
>> <http://www.w3.org/community/schemabibex/wiki/Use_Cases#Use_case:_Describe_a_
>> FRBR_Item_.28e.g._a_single_identifiable_book.29>
>> >
>> > ·1.5Use case: Describe a FRBR Manifestation
>> > 
>> <http://www.w3.org/community/schemabibex/wiki/Use_Cases#Use_case:_Describe_a_
>> FRBR_Manifestation>
>> >
>> > Another way to ask my question, is if we use isbd:hasExpression and
>> > isbd:hasItem [1], does that address these problems? If so, then it's a
>> > way forward.
>> >
>> > GD: The two isbd properties exist only in my head at the moment. The
>> > problem of interoperating data based on FRBR WEMI with data based on
>> > ³dumber², high-level models based on Resource/CreativeWork is generic
>> > (with ISBD, Schema, and DCT as the most prominent examples). So it would
>> > be good to get a general solution.
>> >
>> > Gordon, if this sounds dense it's probably because I need more context
>> > on the ISBD Review Group's work--beyond ISBD identifiers they appear to
>> > have envisioned FRBR-ER identifiers which are news to me.
>> >
>> > The context of the Review Group¹s proposed work is interoperability of
>> > IFLA bibliographic standards, which include ISBD and FRBR/FRAD/FRSAD.
>> > The Group is also working with JSC/RDA on alignments between ISBD and
>> > RDA elements and semantic mappings between them, and between ISBD and
>> > RDA content and carrier designators. More information can be found on
>> > the IFLA and JSC websites, but let me know if you have further questions J
>> >
>> > -Jodi
>> >
>> > [1]
>> >
>> > isbd:hasExpression rdfs:label "has expression" ;
>> >
>> > rdfs:domain isbd:C2001 ;
>> >
>> > rdfs:range frbrer:C1002 .
>> >
>> > frbrer:C1002 rdfs:label ³Expression² .
>> >
>> > isbd:hasItem rdfs:label "has item" ;
>> >
>> > rdfs:domain isbd:C2001 ;
>> >
>> > rdfs:range frbrer:C1004 .
>> >
>> > frbrer:C1004 rdfs:label ³Item² .
>> >
>> > -Jodi
>> >
> 
> --
> Karen Coyle
> kcoyle@kcoyle.net http://kcoyle.net
> ph: 1-510-540-7596
> m: 1-510-435-8234
> skype: kcoylenet
> 
> 
>