W3C home > Mailing lists > Public > w3c-wai-er-ig@w3.org > October 2001

Re: log of chat: EARL and TDPL

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sat, 20 Oct 2001 01:37:14 +0100
Message-ID: <055701c158ff$7a70c480$7ce793c3@y0r1d9>
To: "Jim Ley" <jim@jibbering.com>
Cc: <w3c-wai-er-ig@w3.org>
> [...] this has led to me wasting lots of my time, I hope
> none of EARL has been wasted.

I think that's unavoidable in any project of this nature :-) But on the
whole I think we've been doing fairly well... 7/10? All input is welcome
though, and we received only scraps of feedback for nearly a year now... I
know that some people have questioned various aspects of EARL development,
and yet I am also aware that the feedback from those people has been low. I
can't understand why that is so!

> > 22:17:36 <sbp> well, for a start: bug with 1.0: how do we
> > make conformance claims about a whole site?

A couple of notes:-

a) I meant "0.95".
b) I didn't really mean bug: as I noted, it can be done, we just haven't
really settled the technology at that layer yet. Side-effect of the top
down approach? That's my excuse, and I'm sticking to it :-)

> Definition of site needed?

Clearly. The problem as ever is deciding how much power we want EARL to
have. It could have incredible power, stating that all pages from a site
with a certain extension in a certain folder are excluded, but that all
others matching a certain RegExp are included, with priority levels and so
on... there's always the question of how far we go, trading of
expressiveness of the language for the sake of implementation. Of course,
we also strive to make it extensible, and although it clearly is, it's
difficult at this stage of the evolution of RDF to settle things down.

> By startsWith, I assume you mean that to be a url fragment? so
> log:startsWith "http://example.org/chickens/" would refer to
> http://example.org/chickens/bantam.html but not
> http://example.org/dogs/  ?


> My own thoughts were an EARL "testPage" that just referred
> to a "template" or "sample" or "class" or something I don't know
> the word, but this class of testpage could then be referred to
> throughout the site and linked to with a LINK element.

What you're doing in essence there is "typing" the page, which is fine. You
don't even have to have that information in the page; you can have it,
annotation style, in a third party database. As usual, there are about 5
different ways in which one could approach the situation, which acceptable
results. It needs someone to weigh them all up against each other and say
"this is how it's going to be".

The problem with link types is that HTML is messy in that department. So
I'd prefer the remote typing solution, but perhaps that wouldn't be as
viable to people who just want to go into their pages and say "this one is
marked"? But then, it might be more sensible to try to get them to do it
the other way.

Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Friday, 19 October 2001 20:38:21 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:01:33 UTC