QNames, Ntriples, and Formal Semantics

I am a definite newcomer to this area (having just had RDF be the topic of the week in the final week of an 8-week course on Web Applications, starting from XML in week 1).  It was a great crutch for digging into something I had kept putting off.  I feel relieved that I didn't have to work through that particular history around QNames.

As a neophyte, I have the privilege of noticing things that might have become invisible, so I don't hesitate to speak up.

As long as I'm here, I thought I'd point out some things in the new working documents that I found very valuable and useful to me as a newcomer level.

[Oops, this is not the discussion list that I thought it was.  Oh well, onward ...]

1.	The many primer statements cautioning against overgeneralization from what something signifies to us to what the machine is informed of is extremely valuable.  I find great mischief in the popular overloading of "meaning," even moreso than the over-working of "intelligence."  I am heartened by the effort to be clear about the difference between preserving meaning/interpretation and defining meaning in some colloquial sense.

2.	The formal semantics work is very helpful, especially the model-theoretic approach.  I don't get all of it, by any means.  I will have to invest careful, meticulous reading.  I expect that to be rewarding from what I have seen so far.  I always find it difficult that we are confined to using language to talk about language, and it is easy for me to get lost.  (For example, I think I found some places where the word class is used and something like class denotation is meant, but I am too shaky on the fine points to be certain.)

3.	Then there are Ntriples.  I don't think I could have gotten anywhere if it hadn't been for Ntriples.  What I notice, though, is that Ntriples are too simple and I started using @prefix at once (same reason as those entity definitions ?:).  I also found myself using informal notations to distinguish cases that seem to be lumped under blank nodes -- ways to help me reconcile defined-and-used only-here (the private single-occurrence case), defined-and-used-only-locally (internal to the set), and vocabulary created here (and usable from elsewhere once a URI for "here" is established), as opposed to vocabulary grounded/anchored elsewhere.  And I see I am already in trouble using terms of location (which "here" do I mean exactly - I said I get lost in languaging about language, yes?).  Sigh.  Still, Ntriples are a great help.  And the advice about not over-using a URI for what it is in some near-in sense (e.g., a mailto:...) for the persona associated with that entity is very useful.  

4.	Not using verbs as the local name part of property QNames.  Thank you, thank you, thank you.  The confusion of predicate in RDF with predicate in schoolroom English-sentence diagramming is seriously not helpful and I am gratified that it is not in the material I've digested here so far.

Those are some aspects that I find encouraging and helpful.  Maybe this is better as a back-channel conversation, but I thought it would be good to say publicly what I see working for one person wrestling with these materials for the first time.

-- Dennis

-- orcmid operating on Compagno
-------------------------------
Dennis E. Hamilton | on-line M.Sc in IT student | KIT eLearning
mailto:Dennis_Hamilton@embanet.com | gsm:+1-206.779.9430
mailto:Dennis.Hamilton@acm.org | http://orcmid.com 
OpenPGP public key fingerprint BFE5 EFB8 CB51 8781 5274  C056 D80D 0C3F A393 27EC

-----Original Message-----
From: Frank Manola [mailto:fmanola@acm.org]
Sent: Friday, October 03, 2003 11:16
To: pat hayes
Cc: dennis.hamilton@acm.org; w3c-rdfcore-wg@w3.org
Subject: Re: Internal DTD Examples Invalidate the RDF/XML Documents


pat hayes wrote:

> 
>>  I think that question is already on the table, however.  I'd be 
>> surprised to see such a dramatic change to RDF at this point.  The 
>> QName-URI mapping game seems pretty bolted in.  -- [;<)
> 
> 
> I think the phrase that was used in the past was, the QName ship has 
> already sailed out of the harbor.
> 


And, as history teaches us, if you approach a QName ship, you'd better 
be damn careful :-)

--Frank

Received on Saturday, 4 October 2003 01:33:21 UTC