See also: IRC log
-> http://www.w3.org/2015/05/05-auto-minutes.html Previous meeting
paul: for agenda we have recap of
Stuttgart, test framework, security TF
... please fill in the survey regarding f2f
... github repo has been cleaned up
<kaz> Agenda: https://lists.w3.org/Archives/Member/member-automotive/2015May/0022.html
paul: we had a call for editors
and it has been answered
... editors: wonsuk lee, qing an, adam crofts, kevin gavigan,
and pending is justin
... we also discussed forming a security and privacy tf which
ted reported previously was being held up
ted: resolved internal concern and hoped to have a draft cfp for today but do not have it yet. i will produce by eob tomorrow
<kaz> https://lists.w3.org/Archives/Member/member-automotive/2015May/0022.html
paul: kaz, you wanted to report on testing
kaz: i got some feedback on webplatform testing from others like html5 wg practices
<Paul> New approach:
Implementation Report based onWeb Platform Testing, e.g., HTML5
http://w3c.github.io/test-results/html/details.html
Slides on Google Docs by Mike Smith
(PDF version of the slides)
Conventional approach: Implementation Report by hand, e.g.,
SCXML http://www.w3.org/Voice/2013/scxml-irp/
<kaz> kaz's write up
kaz: we should follow html5's
practices
... as we approach candidate recommendation phase we'll need
reference implementations and being able to produce such a
report based on test suite would be desireable
<kaz> https://docs.google.com/presentation/d/1IKpt626qpPnMpiFP9kbkTZaQPj9LST6Awa8d9zK7WCU/edit#slide=id.p
kaz: please see mike's slide
deck, starting from page 4
... page 8, web-platform-tests testsuite covers 60 specs and
includes 3700 test files and we have 210,000 test results
captured
... it is github based, uses testharness.js
... it runs on w3c-test.org site directly and one can also
choose to use a local runner
... page 12 gives a sample js of calling testharness
... page 13 showing results
... one can run wptserve to run tests locally on a python web
server
... it can run fully on the server if people prefer
... in summary, w3c provides an online testing framework and
provides people the ability to run their own locally as
well
... my recommendation is to go this direction
<kaz> https://lists.w3.org/Archives/Member/member-automotive/2015May/0024.html
-> http://www.w3.org/Voice/2013/scxml-irp/#tests Sample implementation report
kaz: we can generate this
document manually from the information produced by running the
framework
... we need to clarify all the features based on the spec and
define them as assertions
paul: thank you, any comments or questions?
-> https://github.com/w3c/web-platform-tests other specs' tests
<djensen47> Couldn't get to "unmute" fast enough. No comments yet.
paul: i want people to look at this further
adam: this looks reasonable to me but wonder how people will run this themselves on their platforms
kaz: we'll need to clarify what should be checked
adam: if we're just testing the interface. that might be the minimum needed for w3c rec track but group might want to check data coming back out
paul: you can't test an interface without testing the returns
kaz: it is more for testing the
specification not implementations
... there are people looking to extend this for product
conformance as well
dave: w3c doesn't currently have any framework for testing the api itself?
[kaz, should i see if denis or mike is available? he helped code this and may be able to give more details about extensibility]
<djensen47> mocha and jasmine
wonsuk: device api wg made a
sensor api. do we need to check how they made their test cases
using this platform?
... it would be good to see how they are doing it given the
similarities
<djensen47> http://www.w3.org/2009/dap/
wonsuk: i can take an action item to look into how they tested their api
denis ah-kang, mike smith and robin berjon from w3c
dominique hazel-massieux - re device api tests
<kaz> ACTION: ted to work with Kaz and other W3C Team mates, e.g., Denis, Mike, robin and Dom, on how to use the Web Testing Platform [recorded in http://www.w3.org/2015/05/19-auto-minutes.html#action01]
<trackbot> Created ACTION-7 - Work with kaz and other w3c team mates, e.g., denis, mike, robin and dom, on how to use the web testing platform [on Ted Guild - due 2015-05-26].
trackbot, status
ACTION: wonsuk to find out how Device API WG is handling testing of sensors API [recorded in http://www.w3.org/2015/05/19-auto-minutes.html#action02]<trackbot> Created ACTION-8 - Find out how device api wg is handling testing of sensors api [on WonSuk Lee - due 2015-05-26].
action-8 due 2015-07-01
<trackbot> Set action-8 Find out how device api wg is handling testing of sensors api due date to 2015-07-01.
paul: we have been collecting
some comments in github repo on the spec
... some today from peter
<Paul> https://github.com/w3c/automotive/issues
paul: as luck would have it we have all the people who raised comments on the call
dave: for-in is used incorrectly
https://github.com/w3c/automotive/issues/1
dave: it can be fixed easily, see sample code
<kaz> Present now - Paul_Boyes, Adam_Abramski, Greg_Brannon, Dave_Jensen, Kevin_Gavigan, Kaz, Peter_Winzell, Qing_An, Ted_Guild, Wonsu_Lee
paul: as an administrative question should people submit pull requests?
ted: sure, first editors should add themselves as editors on the specs themselves. group created so they can handle pr
dave: the other i started
discussion on ml, pointing back to old issue on bg repo about
api design
... the readme doesn't have much content and pointers on how to
get people more involved
... code guidelines, tools used, participation guidelines
... not sure where we want to organize all that, perhaps just a
link to a wiki
wonsuk: it is a good idea
paul: qing an can you go over your comments?
qing:
... another is for the discovery api
https://github.com/w3c/automotive/issues/6
qing: we need to modify the
current use case section for [voip issues]
... i think there are some problems in the zoneposition
interface
Qing An clarified in email after the call:
Peter: we need to be clearer in various areas, is there subscribe/unsubscribe features for callbacks for instance?
scribe: should we specify the data return values as well?
dave: yes and also how the data
is specified
... on the auto bg (issues list) there were two comments made
on the api
... one was about specifying the data returned. how is the data
specified? is it suppose to be an interface or a
dictionary?
https://github.com/w3c/automotive-bg/issues/49
<QingAn> https://github.com/w3c/automotive-bg/issues/49
ted: iirc the reaction kevron had
is the various oem data points are going to vary too
widely
... we had a call with them but not sure it was minuted
paul: i will need to swap back in
on that. we are running out of time on this call
... it might be worth having a breakout call on issues
alone
[scheduling discussion]
paul: i'll send something on timing for a semi-regular friday call, early my time since that is good for other