Re: MGET and machine processing

PatrickS:
[...]
>> We
>
> Who is "we"? Agfa? Some other group?

Myself and Euler :)

>> assume that Web and SW are unified/reaching their potential
>> when we simply follow http://www.w3.org/2000/10/swap/doc/Reach
>> So log:semantics can easily result in an actual HTTP GET of
>> for instance
>>   eg:r1 eg:p1 eg:r2.
>>   eg:r2 eg:p2 eg:r3.
>>   eg:r3 eg:p3 eg:r1.
>> which is just an example of 3 statements about 3 resources
>> but which is not particularly connected to (and so MGET-able
>> from) one of those 3 resources.
>
> I'm sorry, Jos, but this seems to completely miss the issue.
>
> I'm fully aware that one can GET an entire RDF document. But while
> a particular document might constitute a concise bounded description
> of a particular resource, it need not (and usually won't) so to
> that log:semantics is completely useless.
>
> I'm really unsure what point you were trying to make here. If
> you mean that log:semantics and document-based interchange
> provide the SW, then I strongly disagree.

My opinion is not important. I'm just testing
that design and haven't experienced an issue
observable as a test case.

> They are useful components of knowledge interchange,

That seems a fair aknowledge

> but provide
> far, far too coarse a granularity for efficient resource-centric
> knowledge discovery.

I'm trying to understand what you mean by that...
Of course the (re)source (meta)data could live
in rdbms's, triple stores, etc. The use of the word
``document'' is more in the sense of an RDF graph
and an RDF formula imaginable as being something
written on a separate and discrete sheet of paper.
Anyhow, Amen to your conclusion :)

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/