Re: pfps-16, proposed resolution

At 18:46 01/04/2003 +0100, Graham Klyne wrote:

>With reference to the issue:
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-16
>   "RDF's Expressive Power"
>
>I propose that this is addressed by revisions per:
>   http://www.ninebynine.org/wip/RDF-concepts/20030401/Overview.html
>
>Specifically:
>
>(1) deletion of old section 2.2.7 (which didn't really say anything not 
>covered by 2.2.6)

fine


>(2) revised title and rewording of section 2.2.6:
> 
>http://www.ninebynine.org/wip/RDF-concepts/20030401/Overview.html#section-anyone

I see this as in part a response to pfps concerns about "say anything ..." 
and related to pfps-15.
It does look like an editorial change to me and worthwhile if it assists 
pfps in accepting the disposition of pfps-15.

I have a problem with the sentence:

[[
It is not a goal that RDF cannot prevent anyone from making assertions that 
are nonsensical or inconsistent with other statements, or the world as 
people see it, and applications that build upon RDF are expected to find 
ways to deal with incomplete and conflicting sources of information.
]]

The complexity of the sentence, the double negative at the start bother 
me.  So does the last bit, though I accept it was in the lc wd.  My 
applications do nothing about dealing with incomplete and conflicting 
sources of information.

I suggest:

[[
RDF can be used to express apparently nonsensical or inconsistent 
statements.  It is not the role of RDF to try to prevent the expression of 
such statements.  Designers of applications that use RDF should be aware of 
this and may design their applications to tolerate incomplete or 
inconsistent sources of information.
]]


>(3) revised content of section 3.5:
> 
>http://www.ninebynine.org/wip/RDF-concepts/20030401/Overview.html#section-SimpleFacts

Overall, I think this is an improvement.  I like dropping the example use 
of b-nodes and referring to the primer.  I'm tempted to wordsmith, not sure 
I should:

[[A simple fact expressed in RDF may indicate a relationship between two 
resources, in the form of an RDF triple in which the property names the 
relationship, and the subject and object denote the two resources.]]

trying ...

[[A simple fact expresses a relationship between two resources.  A simple 
fact is expressed in RDF using a triple.  The predicate of the triple names 
the relationship and the subject and object name the two related resources.]]

Hmmm, ... not sure that's any better ...

[[ i.e. assertions of named properties about identified things.]]

Do we still talk about assertions?

I suggest dropping the last two paragraphs.  These get us into murky 
waters.  They are not needed to dispose of macgregor-01/02.  Better, I 
suggest, not to mention them.

Brian

Received on Wednesday, 2 April 2003 09:35:15 UTC