Re: Tim Berners-Lee on the Semantic Web

From: Jonathan Rochkind <rochkind_at_nyob>
Date: Wed, 21 Oct 2009 10:35:41 -0400
To: NGC4LIB_at_LISTSERV.ND.EDU
I wonder how you can say "it's not FRBR".  Pretty much all FRBR says is 
"cluster manifestations into works". Well, it suggests clustering them 
into expressions too, but I don't think not doing that makes it "not 
FRBR".  Nor do I think how you choose to cluster manifestations into 
works can make it yes or no FRBR; the FRBR model actually offers very 
little guidance on HOW to cluster manifestations into works, it just 
says "see, there's this thing called a work, and this thing called a 
manifestation (and this thing called an expression), and when we record 
data we should keep that in mind."  (And really, I think it all makes a 
LOT more sense if you consider those 'things', work, expression, and 
manifestation as sets of items.  The FRBR document doesn't make that 
clear, but neither is it incompatible with that conception).

I think FRBR is both more and less than people seem to be assuming. It 
is simply a domain model for the bibliographic universe, which 
establishes that there IS a thing called a work, and it matters.

In that sense,  LibraryThing and Kindle are both "more FRBR" than the 
typical library catalog.

I agree with Tim that time is a-wasting, it's time to start _doing_.  
Which is why I'm perplexed by the argument that RDA is damned by it's 
association with FRBR, and it's better to just make some changes to an 
AACR2+ that continues as rules for text without any domain model at all. 

You've got to start somewhere, and start doing.  And you NEED a domain 
model.  Starting with the domain model that was formalized from 100 
years of cataloging tradition seems as reasonable decision as any to me 
-- even if it is un-tested, what else have you got that's tested?  It's 
time to start testing it and start doing -- not fiddle around with rules 
for writing text instead of rules for encoding data according to a 
domain model.  Time to move to rules to encoding data according to a 
domain model, and I don't see any better alternatives for this domain 
model being offered than FRBR. (Now RDA... RDA may have other problems 
that are NOT related to it's attempted basis on FRBR model.  That basis 
is one of it's few unassailable positives to me.) 

If you want to just continue fiddling with rules for writing text 
without having a formal domain model... well, like Tim says, put your 
data out there _anyway_, and maybe someone will find a way to do 
something with it.  But you're sure not making it easy, for any of us, 
inside libraries or without, trying to use the data.

Jonathan

Beacom, Matthew wrote:
> Tim, 
>
> You said, "The Kindle store has a work-clustering concept too. And it's not FRBR."
>
> I wonder if you could tell us some more about the Kindle store's work-clustering concept? I ask because I wonder how different from the generic outline of WEMI in the FRBR any work concept relating to textual works could be. After all, FRBR is largely a re-statement in fairly rigorous and formal terms of the traditional publication concepts and practices relating to the idea of a "work," the idea of "editions," and the practices of book production itself that create sets of individual units that contain the "work." I guess my question comes down to this. Is the Kindle work-clustering concept based on traditional publishing concepts and practices coded into a useful algorithm or is the Kindle work clustering concept based on some other more or less formal and rigorous articulation of either the traditional practices and ideas or some innovative alternative?
>
> Matthew Beacom
>  
>
>   
Received on Wed Oct 21 2009 - 10:39:48 EDT