Re: A Crack at an EARL Vocabulary

On Wed, 7 Feb 2001, Sean B. Palmer wrote:

  > I figured that a detail is different from a comment in that it is
  > attached to something at a different level, not that it is a different
  > type of property?

  I think a detail is something that is more specific, like an assertion,
  whereas a comment is simply generic - a bit of prose. I think earl:asserts
  deprecates earl:detail.

OK, then I don't understand again. The scheme I had allowed assertions to be
made at arbitrary levels of detail - you could make an assertion about a
namespace (given a suitable URI for a namespace) conforming to a
specification, or you could make an assertion about anything addressabele (a
text range, using Xpointer) meeting a particular requirement that hda a URI -
for example a technique in AERT. If that's what you mean then I think we are
better off doing it like that. Otherwise I still don't understand what you
mean... (unless it is that we don't need earl:detail)

  > > Call it a group of pages.
  > Isn't there an RDF construct for this already?

  I don't think so... unless you mean like a "bag", but what if you don't
  want to list all of the URIs? You just want to say "anything uner this
  diman/subdirectory"; a bit like a namespace prefix...

Found it in the Model and Syntax spec at http://www.w3.org/TR/REC-rdf-syntax/

in section 3.4 COntainers Defined by a URI Pattern

<rdf:Description aboutEachPrefix="http://foo.org/doc">
  <s:Copyright> 1998, The Foo Organization</s:Copyright>
</rdf:Description>

(It was to match a feature of PICS I believe)

  > >   earl:result (x has the result y)
  > >   earl:status (x has the status y)
  >
  > Nothing is final and definitive.It's just as far as we got for now. So
  > I think we only need one of these.

  O.K, in that case I'd go for result, because it is a bit more specific.

Sure.

  > (Anyone tracking what properties we still have?)

  Nope :-) I'll make an RDF Schema of all of the properties that we end up
  with when the discussions reach some "satisfactory point".

Hmm. It would be handy to have a list again of the ones we think are useful.

  > While I am at it, there is a question of whether it is helpful to have
  > the three conformance level properties that I had for ATAG relative
  > priorities. They allow us to directly use WCAG as an object, but I
  > am not sure how important each of those goals are.

  I think it is better to use them as objects rather than something that is
  written into a property. Just an opinion, but I think that way you lessen
  the amount of properties that use use, and the system is less complex.

Yes, I agree with you. I just need to figure out how to use taht in relation
to the relative priority stuff of ATAG, where you need to know at what level
something meets a particular checkpoint. I guess we can do it like

X :meets [atag:levelA atag:checkpointN] .

or something.

cheers

Chaals

Received on Tuesday, 6 February 2001 22:15:11 UTC