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

Re: Disjointedness of FRBR classes

From: Karen Coyle <kcoyle@kcoyle.net>
Date: Sun, 30 Oct 2011 12:28:25 -0700
Message-ID: <20111030122825.21044qax4q4cdx7d@kcoyle.net>
To: public-lld@w3.org
Quoting Jon Phipps <jphipps@madcreek.com>:

> Karen, I don't believe any of the current FRBR models support this notion
> of 'statement bundles' (distinct from WEMI entities) very well, and it's
> not clear to me at the moment how best to model it in RDFS/OWL, although I
> think Tom's suggestion to use 'named graphs' in some way is worthwhile. I'm
> not convinced that FRBR properties represented as 'shareable units' is
> workable in the Open World context of Linked Data and is liable to be more
> useful in system-local metadata creation/maintenance strategies rather than
> distribution/aggregation.

Jon, I wasn't thinking of the "bundle v. entity" aspect of sharing,  
but was responding to the idea that the entities bundle re-usable  
portions of bibliographic data (without saying what form that bundle  
takes). FRBR is touted in the library world as being a more efficient  
sharing of library data because it will eliminate redundancy (and  
ambiguity) in library data, and save time. The idea being that once a  
Work has been defined in some large pool of bibliographic data, no one  
else needs to define that Work, they just use the Work entity that is  
there. The Work/Expression combo ditto. This structure is also seen as  
some as creating a user discovery path, either bottom up (MEW) or top  
down (WEM). Many see this as an XML-like structure (probably because  
it looks like a hierarchy), but FRBR itself (the document, not FRBRer)  
is silent on data serialization.

If you state that you have a thing called a Work and it has a certain  
set of "attributes" (the term used in the FRBR document), and another  
thing called Expression with its set of attributes, you do seem to be  
implying that each set is separate (disjoint) from the others. If you  
allow attributes to be shifted among sets, then you have something  
different from what FRBR has defined, I believe. I suspect that if you  
accept FRBR as written, then you have to accept that the entities are  
disjoint. I'm not saying that I favor that, but that's how I read it.

So then the question becomes: are there useful groupings of properties  
that do not require "disjointedness"? Could different bibliographic  
data creators share Works that have different sets of properties? And  
that brings me to: what is the role of the WEMI structure? Is it a  
helpful thought model, or is it a data structure?

kc

>
> Jon
>
> I check email just a couple of times daily; to reach me sooner, click here:
> http://awayfind.com/jonphipps
>
>
> On Sat, Oct 29, 2011 at 2:16 PM, Karen Coyle <kcoyle@kcoyle.net> wrote:
>
>> Tom, this utility of FRBR for bundling statements into sharable units is
>> indeed a key part of its design. The sharable units are not, however,
>> W|E|M|I, but W|WE|WEM|WEMI. I did a blog post on this:
>>   
>> http://kcoyle.blogspot.com/**2010/05/frbr-and-sharability.**html<http://kcoyle.blogspot.com/2010/05/frbr-and-sharability.html>
>> You can see this in the cardinality constraint statements in FRBRer, where
>> an E requires one and only one W, etc.
>>
>> It seems to me that the Tillett/Murray paper (which I don't have a link to
>> at the moment, sorry, maybe someone can provide - I'm on the road) doesn't
>> address this aspect of FRBR.
>>
>> More, perhaps, when I get better Internet access, but I've done about
>> 2-dozen blog posts on FRBR, so there may be more fodder there under that
>> tag.
>>
>>
>> kc
>>
>> Quoting Tom Baker <tbaker@tbaker.de>:
>>
>>  On Wed, Oct 26, 2011 at 09:30:28PM +0100, Ross Singer wrote:
>>>
>>>> > Would it help simply to stop saying that the WEMI classes are disjoint?
>>>>
>>>> What's the benefit in this?  What have I said if I say that this thing
>>>> is a Work and an Expression (since neither of these exist in nature)?
>>>>
>>>
>>> If Work and Expression were considered "points of view", at different
>>> levels of
>>> abstraction, for describing a resource, then the utility of FRBR could
>>> lie in
>>> the way FRBR prescribes conventions for bundling particular sets of
>>> statements
>>> about a resource into separate graphs.  If the bundle of statements
>>> conventionally made for Works, and the bundle of statements
>>> conventionally made
>>> for Expressions, were followed with reasonable consistency, they could
>>> help
>>> distribute the maintenance of the set of information held in legacy
>>> catalog
>>> records to multiple agencies.  This is a _practical_ benefit.
>>>
>>> Turning the question around: If Works and Expressions do not exist in
>>> nature,
>>> what is the benefit in saying that a resource _cannot_ be both?   And
>>> what is the
>>> cost?
>>>
>>>  If anything, removing the restrictions seems to dilute the point of
>>>> bothering with FRBR at all.
>>>>
>>>
>>> Removing formal-semantic "restrictions" does not necessarily mean removing
>>> semantic "conventions", which may have real practical utility (see above).
>>>
>>> Tom
>>>
>>> --
>>> Tom Baker <tom@tombaker.org>
>>>
>>>
>>>
>>
>>
>> --
>> Karen Coyle
>> kcoyle@kcoyle.net http://kcoyle.net
>> ph: 1-510-540-7596
>> m: 1-510-435-8234
>> skype: kcoylenet
>>
>>
>>
>



-- 
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
Received on Sunday, 30 October 2011 19:28:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 30 October 2011 19:28:56 GMT