- From: Graham Klyne <GK@ninebynine.org>
- Date: Sun, 21 Dec 2003 08:14:31 +0000
- To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Cc: www-rdf-interest@w3.org
At 21:48 20/12/03 +0000, Jeremy Carroll wrote:
>Graham Klyne wrote:
>
>
>>Or, to put it another way, what do you do with:
>>ID => { < a, b, c >
>> < a1, b1, c1 >
>> < a2, b2, c2 > }
>>ID => { < a3, b3, c3 > }
>>?
>
>
>Bottom, false, 0 = 1.
>
>It's just wrong.
OK, I think that's a reasonable position, just not the only possible
such. And it does allow one to say that a syntactic occurrence of a graph
is closed, or complete.
I took a slightly different line:
If one asserts that ID denotes a truth, then one might reasonably take a
view that all of the graphs associated (syntactically) with ID are true,
which means that ID itself would denote a merge of the individual graphs
(or a union if you were to decide that bnodes can be shared between such
instances within a single document).
I recognize (more clearly now than when I posted my earlier comments) that
this is a *choice* that happens to closely reflects the quads approach to
representing "context"s (and in my approach is suggested by a desire to
remain broadly compatible with the CWM approach to contexts).
#g
------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact
Received on Sunday, 21 December 2003 13:35:46 UTC