W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > November 2002

Re: Syntax doc comments

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Fri, 01 Nov 2002 12:12:22 +0000
To: Jeremy Carroll <jjc@hpl.hp.com>
cc: w3c-rdfcore-wg <w3c-rdfcore-wg@w3.org>
Message-ID: <4924.1036152742@hoth.ilrt.bris.ac.uk>

>>>Jeremy Carroll said:
> <<
> 7.2.1 Grammar start
> So should I say more or less?
> >>
> I prefer less.
> 
> <<
> [[We never disallowed rdf:nil did we?]]
> 
> We didn't micro-decide everything, I asked one, got no replies so
> made a choice.  rdf:nil is a sentinel, we can either:
>    1) not encourage its use as a class or property and forbid it everywhere
>    2) not care, and allow it everywhere.
> 
> Do you want to change to 2) ?
> >>
> 
> 2) seems to be more consistent - it is only syntactic terms like
> rdf:Description that cannot be used as a property; I don't see why using
> rdf:nil as a class is any more wrong than using rdf:subject as one.


Done - I took 2) rdf:nil not special (I got this in other comments too).


> 7.2.18
> <<
> MUST
> >>
> 
> My problem with the MUST was not what you were trying to say, but what you
> actually said.
> 
> What you were trying to say is (more or less) that the grammar treats
> "FooBar" like "Literal"; what you actually said is that processing MUST
> (i.e. all implementations have to do this way) continue at a rule which some
> implementations won't even have. (e.g. SNAIL does not follow this framework
> at all; a conventially parser like VRP works over strings and so cannot be
> said to have production parseTypeLiteralPropertyElt and so cannot follow
> this MUST).
> 
> Yes you have carefully said that implementations only have to have the same
> effect, that observation makes the MUST incorrect. This is really linked to
> my observation that the MUST, SHOULD etc terminology is not used
> sufficiently in this doc to justify its inclusion.
> 
> It is possible to rephrase in a way that does not use this MUST,
> 
> cf. Dan's
> http://www.w3.org/2001/01/mp23
> [[
> I try to use the word MUST to constrain agents in processes, not to just
> make declarative statements; i.e. I think it's a misuse of RFC2119 to say
> things like "2 + 2 MUST be 4"
> ]]

Again you offer no actual words to use so I've changed from

  Processing MUST continue at production *parseTypeLiteralPropertyElt*.

to
   This production matches and acts as if production
   *parseTypeLiteralPropertyElt* was matched.

which I hope also has the benefit of removing the word 'processing'.

Dave
Received on Friday, 1 November 2002 07:13:53 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:53:55 EDT