W3C home > Mailing lists > Public > public-grddl-wg@w3.org > February 2007

Re: "GRDDL Agents" section, normative security text for review (# issue-conformance-labels)

From: Dan Connolly <connolly@w3.org>
Date: Mon, 19 Feb 2007 15:44:21 -0500
Message-Id: <1c620c379259ade5390b7c11ba0e331d@w3.org>
Cc: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "GRDDL Working Group" <public-grddl-wg@w3.org>
To: "McBride, Brian" <brian.mcbride@hp.com>

On Feb 19, 2007, at 9:41 AM, McBride, Brian wrote:
> Reading
>
>   http://www.w3.org/2004/01/rdxh/spec [editor's draft $Date: 2007/02/17
> 21:58:04 $]
>
> Section 7 begins:
>
> [[
> A GRDDL Agent is a software module that computes GRDDL results of
> information resources. Given a URI I of an information resource IR, and
> an XPath node N for a representation of IR, subject to security
> considerations below and local policy and configuration, a GRDDL Agent
> should:
> ]]
>
> I think this is ambiguous.  It could be saying:
>
> Forall ?X, if ?X computes GRDDL results of information resources then 
> ?X
> is a GRDDL Agent
>
> And therefore, if anything computes any GRDDL result it should conform
> the rules in the list that follows.
>
> It could be saying:
>
> If ?X is a GRDDL agent then ?X computes GRDDL results of information
> resources and should conform to the rules in the following list.
>
> I think you probably mean the latter.

Hmm... I think I said what I meant; when elaborated
using if/then, it becomes...
   ?X is a GRDDL Agent if and only if ?X computes GRDDL results
   of information resources; if ?X is a GRDDL Agent then ?X should...

But I'm not at all sure that's more clear than the usual
"a triangle is a three-sided polygon" style of definition.


> I note that section 8 is written in terms, not of a GRDDL Agent but of
> GRDDL-aware agents.  So I'm not sure if the intent is to keep this term
> also in the document.

I don't think so; I sorta ran out of editing time and left it there to
see if anybody would notice. I think it's best to scrub "GRDDL-aware
agent" and go with "GRDDL agent" exclusively, but I'm interested in 
other opinions;
in particular, should the use cases document use "GRDDL agent" as well?

Somehow "GRDDL-aware agent" seems like a good choice of words if
we're not really introducing a new conformance label, but as long as we
are, we might as well strike "aware" and go with "GRDDL agent".

>   If so, it might be helpful to distinguish it
> clearly from a GRDDL Agent.
>
> Does the following express what you mean?
>
> [[
> A GRDDL Aware Agent computes GRDDL results according to the rules in
> this specification

That suggests that the rules say how to compute the results rather than
just saying what graphs are [correct] GRDDL results. I don't want to 
suggest that.

>  but may be designed not to compute all the GRDDL
> results of an input representation.

As an elaboration of "local configuration," that seems not quite
useful without giving an example or otherwise saying *why* it
might be designed that way. I'm inclined to let "local configuration"
stand on its own.

>   A GRDDL Agent is a GRDDL Aware
> Agent that conforms to the following requirements.  Given a URI I of an
> information resource IR, and an Xpath node N for a representation of 
> IR,
> except where proscribed by the security considerations described below
> or by local policy as expressed in its configuration, a GRDDL Agent
> should:
>
> ...
> ]]
>
> Or am I just confused?

I don't see any evidence of confusion; our stylistic preferences
seem to differ, though.


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Monday, 19 February 2007 20:44:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:47 GMT