HTML editors

timeline

Chaals: goal is to ship a REC within this charter
... which means PR in August
... and CR by mid-June to run the patent clock
... so 2 variables
... do we get enough stuff done by then?
... and do we get consensus from the WG?
... the approach is:
... lots of changes: ficaption, summary, etc.
... things that didn't make into HTML 5
... we'd like to improve editorially as much as we can

Steve: can we change tyhings during CR period?

Chaals: yes, but we assume we'll have a branch
... and we'll leave it alone at that point
... and do changes on 5.2

Steve: so CR by mid-June. this gives us 2 months

Chaals: yes
... we won't get everything we'd like by then

Steve: when do we identify things to be pulled out?

Chaals: I'm going to start later tonight Call for Consensus
... but if things don't work, they should go
... like outline stuff
... accesskey
... this dooesn't work

Steve: accesskey attribute?

Chaals: no the changes between html 4 and html 5 are all fantasy
... html 5 stuff should be reverted

Steve: re outline algo
... it's no problem sitting since he doesn't have requirements
... but it needs better explanation
... and needs to be broken out
... [...]

Chaals: if it's an element that never got implemented, we just pull it out
... we might need call for consensus for some
... substantive changes should be justified
... and be accompanied by tests

Alex: what about updates to the WD?

Chaals: montly
... next update is next week

PLH: WOuld like to take the snapshot Friday for publication

CMN: Seems fair, I'll shoot for Friday then.

Plh: I'll take whatever is in the repo on Friday then

Interop testing

Chaals: approach is that we want tests to justify changes
... HTML 5.0 got through an interop test
... if all the changes that we make come with tests, that should be a no brainer to go back to the Director
... while I would like to have tests for the testharness, it's not a requirement
... just a nice to have
... so manual tests are fine
... in the long term, we'll keep doing better

Leonie: is there any template or structure for tests?

Chaals: yes
... testharness

Plh: I'm happy to help

<scribe> ACTION: plh to organize a testharness teaching session [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action01]

Chaals: anounce it to public-html

Plh: I'll organize a doodle poll

Charles: does the testing plan make sense to you?

Plh: yes, as long as we cut out features that are not implemented enough

Charles: agreed. nothing should stop us from being ruthless

Leonie: do we have a list of things to remove?

Charles: I have 4

Travis: +1 to cut

Alex: what's our bar? 2 browsers

Steve: are we going to vary from the HTML5 exit criteria?
... it was 2 rendering engines

Travis: it was passive permissive
... because it was widely deployed

Chaals: other parts have not been tested and were never implemented
... so a c ouple of engines
... but there is no general rule at W3C
... having implementers swearing they'll do it might be convincing

<SteveF> exit criteria for html5 https://dev.w3.org/html5/decision-policy/public-permissive-exit-criteria.html

Chaals: but having implementers saying they'd like to remove it, might make it the other way

<chaals> "is sufficiently clear, complete, and relevant to market needs, to ensure that independent interoperable implementations of each feature of the specification will be realized"

Steve: we need to have those rules written down
... why not just reused the ones from HTML5?
... and do a Call for Consensus to use that again

Plh: to give a little background on the criteria
... the goal was to use passive permissive for old html4 features, like tables
... and make sure we had tests for the new HTML 5.0 features
... but some things fell into cracks, like mediaGroup
... we would have never shipped mediaGroup in HTML 5.0 if we had realized it wasn't implemented.

<Zakim> chaals, you wanted to say yeah, let's make something based on those criteria

<LJWatson> +1 to public permissive v3 exit criteria

Chaals: If we make these requirements part of the acceptance for each change, then when we enter CR, we don't need to spend a month writing a full test report

Steve: we've got to have something to define our criteria
... because people we'll be asking questions
... reusing the ones from HTML 5.0 means the work is already done for us
... if we spit up the spec and flag things up for removal...

Arron: I have a tool that can run through the APIs in the spec
... how about I create issues on those who don't have 2 implementations?
... I'll raise issues

[All: +1]

Arron: I hope to have the info in the next 2 or 3 minutes. it's a public tool

<arronei> http://aka.ms/apicatalog/

Chaals: I'll setup a call for consensus to the Group for the exit criteria

<scribe> ACTION: Charles to get the Group to decide on the HTMl exit criteria [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action02]

<chaals> ACTION: Chaals to CfC on "exit criteria" as a requirement for accepting any change [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action03]

<scribe> ACTION: Arron will enter issues based on API implementation feedback [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action04]

Bug Triage

Chaals: we cleaned out the a11y bugs and give action items away
... but there are a bunch of other bugs from bugzilla
... how do we triage github issues?

<LJWatson> http://github.adrianba.net/webstandards/HTML5-bugs.htm

Plh: asking the editors to sort out which bugs are 5.1 would be good

<scribe> ACTION: Plh to create a milesone 5.1 in the repo [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action05]

Travis: what happened to the issues we don't migrate?

Chaals: they'll disapear over time

Travis: incubator group project?

Chaals: yes. the bugs might come back again however

Leonie: let's finish the a11 bugs this week
... and see where we are after that

Arron: we'll look into organizing a bug triage call

Getting up to speed

Chaals: do you feel in control?
... or need help in figuring things out?

Travis: curious how we get the commmunity involved in discussing the issues
... we could divolve into a benevolent editor situation but we should be more inclusive
... how do we say that an issue needs someone to look into it?
... does it jump out of the incubation group?

Chaals: we have a label for question for the WG in github
... and we need to practice that
... once we start doing that, we'll need a tradition of answering :)
... the WhATWG spec is removing a lot of semantic elements

Steve: was an Aprils' fools

Chaals: yes

Steve: to get the Group involved, we'll need to write to them

Plh: weekly report? Twitter?

Travis: I'd like to ask people in the trenches

Plh: we can use the w3c twitter account

Steve: and point people to the github issue

https://twitter.com/htmlwg

Steve: we can build it up

<scribe> ACTION: plh to find who has access to @htmlwg [recorded in http://www.w3.org/2016/04/05-html-minutes.html#action06]

Chaals: linking stuff in bikeshed
... seems black art to me

Arron: I know how to link things

Alex: getting the community involved is important and getting things on github is also good rather than email

Chaals: formal comments need to be put into GitHub

Next meeting

April 26, same time 2000 Zulu