Re: [tangle] getting the semweb exactly wrong

Timothy Falconer wrote:
> On Jan 3, 2006, at 12:35 PM, Frank Manola wrote:
> 
>> Timothy Falconer wrote:
>>
>>> Blog post excerpt:
>>> "Reading such comments confounds me, since they've got it *exactly* 
>>> wrong. The Semantic Web approach is LOOSE, not normalized.
>>> ...
> 
>>  ...
> 
>> RDF data is *highly* normalized: RDF essentially organizes data as 
>> binary relations (one per property) with surrogate keys (URIs), which 
>> is as normalized as you can get.  This high degree of normalization is 
>> one of the things that makes the data structure so flexible. RDF is 
>> looser than the relational model in some other respects, but they have 
>> nothing to do with normalization.  "Normalized" isn't properly the 
>> opposite of "loose" either
> 
> 
> Frank,
> 
> 
> You are of course correct.  As Danny pointed out, I was responding to 
> David's comment.    I did balk at the term "normalized" when I wrote it, 
> and tried a few other terms like "too constrained", "brittle", "rigid", 
> "limiting", etc, but they didn't flow from the quote so left it as is.
> 
> 
> Probably the best word to use in answer to his quote is "un-webby".  
>  RDF is "webby", not "un-webby".    Remember, David Weinberger's the guy 
> who wrote "Small Pieces Loosely Joined", so being webby is a big thing 
> for him, as it is for a lot of us.   Being webby's what made HTML/HTTP 
> take off over the other more prescriptive hypertext schemes of the time.
> 

Timothy--

Good.  Actually, I think "brittle" works pretty well, but searching for 
a single, one-word descriptor for this is likely a losing proposition. 
Look at how well the use of "normalized" worked, for example!  I don't 
really like "un-webby" either, since you wind up having to say what that 
means in more conventional terms.

What's odd, when you think about it, is that the author of a book called 
"Small Pieces Loosely Joined" should object to a Semantic Web based on 
RDF which, after all, involves different individuals describing things 
that are interesting to them by adding "small pieces [triples] loosely 
joined" to the Web.  Can't get much smaller (or more normalized) than a 
triple.

Part of what may be going on here is the frequently-occurring confusion 
that imagines that the use of schemas/ontologies on the Web to describe 
terminology implies that everyone needs to use the *same* terminology. 
What is actually going on, of course, is that people are free to use 
their own terminologies, borrow from others if they wish, or use 
existing terminologies in their entirety.  Using URIs for the terms 
keeps all this straight.  The terms may or may not have definitions in 
schemas or ontologies.  People can come along later and identify 
relationships between those terms, or create (or add to) Web-accessible 
definitions.  Once again, "small pieces loosely joined".

How all this is "un-webby" is beyond me.  Surely Weinberger doesn't 
imagine that all the pages on the existing Web use the same terminology, 
or that the Web can't be useful without a given user being able to 
understand all those pages (as should be clear by now, I haven't read 
the book).

--Frank

Received on Wednesday, 4 January 2006 22:13:59 UTC