See also: IRC log
<shadi> http://www.w3.org/2006/02/mwi-test-charter
SAZ: If anyone is interested in the MWI test description area, please get in touch with Shadi
<shadi> http://lists.w3.org/Archives/Public/public-wai-ert/2006Dec/0001
SAZ: We need a decision on to publish the doc, I suggest a week to review and then vote with an online strawpoll
CR: Fine with me
SAZ: There's a deadline by 21st December until
Jan. So 1 week to review and 1 week to sort out publication.
... Try for publication around the 20th.
... Get your comments in ASAP though please.
<shadi> http://lists.w3.org/Archives/Public/public-wai-ert/2006Dec/0002
SAZ: Seperate out a super class to webcontent and content - Every exchange stored in a seperate instance of the http content class.
JK: Looks Good
JL: Looks fine, but XML validity issues
SAZ: Any one against it?
RESOLUTION: Adopt the content class into the EARL schema
<shadi> http://lists.w3.org/Archives/Public/public-wai-ert/2006Oct/0047
SAZ: Do we need additional properties to hide
and encrypt some data that is in EARL reports
... For example passwords in HTTP class.
... Do we need to do more than make people aware of the issues in the
guide?
CI: I think we should at least highlight the issue, if not include extra properties
SAZ: Are you saying we should do both and add properties as well as note it?
CI: I would be happy with just the highlight note solution, if the group doesn't want to do more.
JK: A note in the guide would be okay, but not so sure about the Schema.
CI: I think it would be important enough for
the Schema.
... could be in the HTTP Vocab
SAZ: I think JK said previously you have the
same security considerations as HTTP
... We can highlight it somewhere there.
... Doesn't need to be too long
... If everyone is in agreement, I'll take an action to get it into the HTTP
Vocab
JL: I don't think we should exclude the idea of having a brief statement warning that there could be sensitive data.
JK: agree
SAZ: Proposal is brief statement in schema and http doc and a small section in the guide any objections?
RESOLUTION: Include brief statement in schema and http vocab, and a small section in guide
<scribe> ACTION: SAZ to include security statement in earl schema [recorded in http://www.w3.org/2006/12/06-er-minutes.html#action01]
<scribe> ACTION: SAZ to include security statement in http vocab [recorded in http://www.w3.org/2006/12/06-er-minutes.html#action02]
<scribe> ACTION: CV and JK to include security section in the earl guide [recorded in http://www.w3.org/2006/12/06-er-minutes.html#action03]
<shadi> http://lists.w3.org/Archives/Public/public-wai-ert/2006Dec/0003
SAZ: There seems to be some agreement that
there's a warning flag, rather than subclassing and creating new values -
have a description or warning that leads to a text description.
... JK had a proposal too?
JK: My proposal was the same as CI's but with subclassing
SAZ: but David didn't like the subclassing
JK: I'm in favour of subclassing, a warning is similar to a test result, but not quite. like a test result only without a validity level
SAZ: With CI's proposal a warning was similar
to test results, and in different frameworks they could easily be test
results instead
... Perhaps a different idea is to have an earl warning property that was to
string
JK: but that would mean you'd have to have test
results for each and every warning, e.g. this is valid CSS but warning X
repeated for every warning
... It would be similar to my proposal, but the validity level would still be
pass, but the warning text would be what I use in the dc:description
... I'm most concerned about what happened with failed tests rather than
passes, so you'd still get failed results and passes
CI: You could have instance in test result or warning
JK: And where would you describe the error
JK: If you had more than one error, where would you have the warnings, in just one of the errors, or all of them, or seperate out to another testResult?
CI: You'd have the warnings in the most appropriate testResult
JK: Would you be able to create another proposal using the same method but when the documents have Errors and warnings rather than just Pass and warnings.
ACTION CI Expand previous proposal to include warnings with failed tests, rather than passed ones.
SAZ: With the examples we have, they could easily be testResults too, so we could have the CSS validates, but note these other results.
JK: Here it's a question of what is the test?
SAZ: So the original test CSS, would be a past,
but there'd be earl warning properties which point to test results too, which
might still be failed tests, it doesn't affect the validity of the original
test
... So in the context of the tests that are being warned it's a fail, but
that doesn't affect the parent result
CI: You don't need to record the links from test results to the warnings, you have 3 test results and 2 warnings
SAZ: My Alternative is to just link tests so even that a particular test succeeds, there are still warnings that these other 2 tests
CI: We'd agreed not to do a test relationship language
SAZ: My concern is that earl:warning is a hidden test result class, and tool developers will just mis-use it rather than use cannotTell etc.
CI: The warning class has been designed to avoid this somewhat by not giving validity levels etc.
SAZ: I'm concerned it will be "valid WCAG 1, but then say earl:warning "check colour contrast"
CI: But that is a misuse, I'm not sure we can do stuff to prevent misuse.
SAZ: It's true we'll never succeed totally, but
this has been a problem in the past, as users like to see a "pass"
... Am I the only one seeing it as a hidden test result.
JK: You could say that, but the point is what is the test?
CI: maybe Shadi could provide an action to how people could misuse a warning as a test result, as I can't see an example
<scribe> ACTION: Shadi to provide misuse of warning class examples [recorded in http://www.w3.org/2006/12/06-er-minutes.html#action04]
<scribe> ACTION: CI Expand previous proposal to include warnings with failed tests, rather than passed ones. [recorded in http://www.w3.org/2006/12/06-er-minutes.html#action05]
SAZ: We should have 2 f2f a year, we should be
thinking of an upcoming one for the beginning of the year
... Perhaps the mini-tech at the end of Jan, but that's a bit soon
perhaps?
... Any date/time preferences?
JK: There was an invitation from Chaals in Norway?
SAZ: and CTIC in spain
... or to St Augustin again?
CI: I like Spain :)
... There's a W3 Symposium in Spain in the first week of Feb.
<CarlosI> http://w3c.es/Eventos/2007/eGov/
<CarlosI> European W3C Symposium on eGovernment
SAZ: Any dates people can't make in Feb?
JK: Only weekends occupied for me
<CarlosI> 1-2 February
SAZ: End of Feb might be best to give everyone time to organise travel
CI: Feb is busy time for CTIC, I will check and find out more
Cheers!