See also: IRC log
<Christophe> http://lists.w3.org/Archives/Public/public-wai-ert-tsdtf/2006Nov/0029
(discussion of allowing only one technique per test sample)
CS: If more than one technique is required for a SC then it means there can be more than one test sample per SC.
CV: Might be rare cases where SC must repeat location of accessibility problem.
CS: Concern is that techniques and locations
are not associated other than one following the other.
... We can likely live without it but may be helpful.
... Several ways to associate techniques and locations. Could be ID
references or nesting.
DO: If nesting then agree with Vangalis other
could use ID references.
... Has test sample where problem occurs.
... Example is form with several labels.
(discussion)
DO: Perhaps not a valid example. Could be dropped.
CV: Relationship between techs and locations is
important. Current strategy is weak and should improve.
... If use nesting then like location inside techniques.
... Always like less verbose strategy. Slightly prefer ID reference.
... Not sure if we can follow ID reference model however.
CS: Can live with ID reference model.
... Asks for examples of each model be sent to list and we can discuss.
<scribe> ACTION: CS will create examples of each model and send to list. [recorded in http://www.w3.org/2006/11/28-tsdtf-minutes.html#action01]
http://www.w3.org/WAI/ER/2006/tests/process
DO: Good structure but in step 3 (intial evaluation) not sure who does assignment?
CS: People may take batches but not sure who
assignes.
... Might be nice to have something in TCDL that states who has been assigned
to test sample.
... Would require a change to TCDL.
CV: Not sure we should include who reviewed test sample in evaluation.
(discussion)
CS: Perhaps use a temporary element in file?
DO: Important that at least one person has overview of who is assigned to test samples.
(use of Wiki discussed)
TB: Has checklist for metrics been created yet?
CS: No checklist yet.
TB: Need common set of metrics.
CS: If sample does not meet all requirements
the returned to 'unconfirmed' status.
... Most of checks are quite clear but there is still some small
interpretation needed.
TB: Could try out and see if works. Then modify if needed.
CS: Tried using test sample in CVS but did not
have all data to move to next step.
... Until we have all data all test samples will be unconfirmed.
DO: In step 4 - "checklist for structure reviews" should this be "checklist for content reviews"?
CS: Yes, should be.
CV: Questions if we need to seperate steps 3 and 4.
CS: Don't see issue with strawpolls.
DO: Suggest we need timeouts for polls.
CS: Yes, can add timeouts and if insufficent responses then can be reopened.
(general agreement)
<CarlosI> sorry, I've lost conection and now Zakim doesn't want me on the conference again
<CarlosI> In brief, I think Steps 2 and 3 are the same as currently defined and could be merged in one
<scribe> ACTION: CS will write up summary of discussion and send to list for comments. [recorded in http://www.w3.org/2006/11/28-tsdtf-minutes.html#action02]
<CarlosI> sorry, I mean steps 3 and 4