See also: IRC log
shadi: changes to draft (shadi+chaals). Added schema in n3, considers what parts should be normative, and dealing with implicit vs explicit matters
chaals: need to clarify which of n3 or rdf+xml is normative in case of trouble
shadi: Is Triples/N3 overdose? Should we just have one?
Chaals: tables are human edited, and triples aren't very useful, complicated to read, N3 is more useful than that.
shadi: Should we have triples or just N3?
JL: Not too much value in having the triples rather than other forms, they're easily generated
Shadi: Should we make triples normative instead?
niq: triples/N3 makes it more complicated for the non XML people
<Zakim> chaals, you wanted to suggest that RDF/XML should be canonical format
Chaals: Drop the triples, put the N3 format in, make the RDF/XML be the canonical version of EARL
Shadi: Everyone agree?
JL: Agree
<Zakim> chaals, you wanted to say no
Shadi: Should we split up the N3 or RDF/XML also into subcategories as we split up the NTriples
chaals: the machine readable formats are for the machine, we've got them split up in the text of the spec, the table of terms makes a good index, but I never look in the DTD or similar, so it should be kept whole for the machine
Shadi: I agree when approaching it from the RDF
side, but this document is a reference, the machine readable part will be in
at the namespace.
... we could add part of the schema at each part where it's defined, or have
it split up in the appendix to find it.
Chaals: People would look at the table not the
schema to get at more than the schema as a whole
... I agree it should be in the schema and reasonably sensibly ordered, but
don't need to spend too much time on it
Shadi: Would a link from terms in the appendix back to the terms definition.
Chaals+JL: YEs
Shadi: would it be good to have a snippet of the schema at the terms definition.
Chaals: Having chunks is over kill, a link to the schema would be enough
Shadi: Should we title and split up the schema?
Chaals: makes sense to do that.
Shadi: Does any body know of a good way of constructing a good node/arc diagram?
Nick: The RDF validator, but that produces very complicated result.
Chaals: By hand is best I'd've thought, IsaViz might be able to do it.
Shadi: Let's add it to the wishlist.
Chaals: It's a lot of work to produce something.
It's something I want to do with foafnaut, and most things are there, but I've not found the time yet.
JL: If we're going to do it by hand we only want to do it once.
Shadi: RDF/XML will be the normative reference
<chaals> RESOLUTION: The RDF / XML is the normative reference
Shadi: Should we make appendicies B/C should not normative aswell, or should we maybe even remove them?
Chaals: Yes we should keep them, but make them non-normative.
Shadi: Should the terms, currently App C, be the first appendix to see as it's first people want to see?
Chaals: Order's not that important.
... A typical W3 spec starts with contents link, then link to the schema and
terms as the most important thing.
Shadi: and the EARL Guide also wants to be up
there easy to find.
... Intro is non-normative, all agree?
[Everyone agrees=]
Shadi: Is it a problem that the table of contents are different?
Chaals: They should be the same yeah, is there a good reason?
Shadi: No good reason currently.
... Everyone should have a closer look at the Spec, and send in as many
bugs/typos/ideas/suggestions to the list?
... Now person with using tool etc.
Chaals: The easy cases become much more complicated, and the hard cases are just the same.
Shadi: The Current schema allows multiple people and multiple tools all together as a single assertor, currently there's always a Person node, is it really that there will always a Person and a tool involved.
Chaals: the usingTool is just about using a
tool to help generate the result of the test, not just the case of generating
the Earl.
... I think there's lots of examples especially outside of WCAG that don't
use a tool to help, UAAG SpecGL are simple person situations.
... We can certainly have cases with multiple tools and potentially no
human.
... Which also means possible to have several people getting together to give
one.
... There's value in multiple tools and multiple users, it's a good thing as
it is now.
Shadi: do we really want to open up the idea that there's clusters of tools/persons/mixes together?
Chaals: We may want to do that?
... One of things we should do, there's a collection of things that an
assertor can be currently, a bag or something so we can extend would be
good.
JL: I like the sound of having the mixtures of people, but if it leads to really complicated RDF, then we should move away from it, good to investigate now.
Shadi: We have persons and tools as in the draft, and I think we should publish this as is now, and keep it open to discuss.
Chaals: We should open it now to a bag straight away, rather than a collection, so as we keep it open so can narrow later, but harder to widen.
Chaals to investigate and report to list
<chaals> ACTION: Chaals look at how to have the Assertor described in a way that can be extended or restricted later. [recorded in http://www.w3.org/2005/07/12-er-minutes.html#action01]
Shadi: Any other comments? Planning to get next weeks version into the queue for publication.
Shadi: Nick are you going to continue on the HTTP headers stuff?
Nick: I forwarded to the list previous stuff.
Shadi: Can you look at the Working Draft and check that the draft is in accordance with your stuff.
Nick: okay
Shadi: Identifying the location of errors is still open - are you free Chris to work on this?
Chris: Sure.
Shadi: We also have the evidence stuff, Jim/Chaals?
Chaals: I have this on my radar, and will look when the draft is out
Shadi: I think we should work in paralell, the
draft is basically stable now.
... The Earl Guide, we need to start working on this
... That's a big job of course, so we need different people working on
different parts.
... Johannes has put his hand up to do part of it already
... Looking for someone to take on the overall editorial side
<chaals> ACTION: everyone, prrofread current draft. [recorded in http://www.w3.org/2005/07/12-er-minutes.html#action02]
Shadi: We should review WCAG 2.0 make sure it's
testable and feedback how testable it is.
... Is it good for tools? Maybe some joint development of test suites.
<scribe> ACTION: everyone read WCAG 2.0 ready for 26th July meeting feedback [recorded in http://www.w3.org/2005/07/12-er-minutes.html#action03]
<ChrisR> test suite for WCAG2: http://www.w3.org/WAI/GL/WCAG20/tests/
Shadi: There's a techniques task force that's
collecting tests, maybe we should bring some of that into this group.
... Earl Requirements Doc is now published
... Thanks all!