- From: Dan Connolly <connolly@w3.org>
- Date: Mon, 19 Feb 2007 23:04:02 -0600
- To: Murray Maloney <murray@muzmo.com>
- Cc: "GRDDL Working Group" <public-grddl-wg@w3.org>
On Feb 19, 2007, at 4:38 PM, Murray Maloney wrote:
> I think that I was the one who coined "GRDDL-aware Agent" and I think
> that
> I should try to explain why I felt that "-aware" was originally
> suggested...
>
> On one hand, we can certainly expect to see a proliferation of
> stand-alone
> agents which meet the criteria laid out by Dan earlier in this thread.
> And I
> can see how "GRDDL Agent" seems much more suitable for that class
> of processor.
>
> On the other hand, and for reasons that I cannot express, I expect to
> see
> GRDDL-awareness being built into future generations of Web Browsers
> -- which are otherwise known as "User Agents." In such cases, I do not
> like to think of the UA as a GRDDL Agent, because it does so much more.
>
> Rather, I think of them as being "GRDDL-aware."
Hmm... that argument suggests that typical web browsers are not HTTP
clients,
since they do so much more. It doesn't appeal to me. The rest of the
argument is more persuasive...
> Further, since I do not believe that such an agent has an obligation
> to perform
> transformations at will, but will rather perform transformations only
> at the
> behest of the user -- who presumably initiates the transformation by
> clicking
> over an anchor or UA button, or by way of user preferences -- I think
> that
> GRDDL-awareness as being critically important to the user experience.
>
> That is, I do not expect to transform every web page that shows up in
> my
> browser window. Rather, I expect that my GRDDL-aware UA will notify
> me when it is aware of GRDDL transformations being available, and
> perhaps
> offer a pop-up or some-such to show me useful triples. In time, my UA
> would
> learn from me which transformations I consider valuable. My UA could
> also
> notify other agents which might use the GRDDL results to populate my
> address book or calendar.
>
> I guess that what I saying that I don't find a GRDDL Agent, in and of
> itself,
> that interesting. Rather, I see GRDDL-awareness being a highly
> desirable
> for Semantic Web applications and agents.
OK, that makes sense... it's a little bit like whether to load images
or not,
if it weren't the norm to load them pretty much every time.
It's possible to squint and look at that as a matter of "local policy
and configuration". That's a stretch... perhaps "At user option"
instead?
Try this on for size?
A GRDDL-aware 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, a GRDDL Agent should, at
user option and subject to security considerations below:
1. Find each transformation ...
>
> Finally, since Dan is a stickler for sticking with decsions, I'll
> point out that
> we already made a decision to use the term "GRDDL-aware agent"
Yes, but we have new information since that decision, so the
chair re-opened the issue 14 Feb, whence comes my action and
this thread.
http://www.w3.org/2004/01/rdxh/spec#issue-conformance-labels
--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Tuesday, 20 February 2007 05:04:07 UTC