dsinger: See the URL https://lists.w3.org/Archives/Public/public-tracking/2017Jul/0007.html for matthias's e-mail
… We need to decide on transition to CR, this week or next.
fielding: We have another week, don't we?
dsinger: But I'm not here next week.
alan: Notify people on mailing list?
… Next week probably also a short meeting, light burden on whoever chairs.
dsinger: If no questions today... Then I'll send e-mail.
mike: I was going to write some stuff after I read Roy's edits.
… Main thing is Roy's idea to rationalize the API.
… I'm busy this week, but can do for next week.
… How do I do that? Edit the spec?
fielding: No, create a Pull Request instead.
mike: OK, a PR, but then not confirm it.
… Simpler API returns same kind of object in all cases.
alan: a question for Roy:
… why do you believe the doc. isn't ready?
fielding: Lacking any real impl. in browsers, and no intention to implem.
… I'm used to no changes after CR.
… I'd prefer the changes I outlines that Mike mentioned and more time to explain them and only then go to CR.
… Fine with publication as WD, but it doesn't seem a CR to me.
… I'm used to specs that are implemented by hundreds of people.
… Find out if they work first.
dsinger: Formally, a CR is a request for implem, but I agree we should have more validation.
mike: Why did you say the JS property is redundant?
fielding: Reality with DNT is you want to find out what headers would be sent to 3rd party.
mike: If you have a piece of JS in a subresource browsing context, you cannot know what the status is, without bouncing it. So a property seems necessary.
fielding: That's the Confirm API.
mike: it seems we should keep it.
fielding: It is still there, it was just my comment.
dsinger: So no decision today. Adjourned.
Succeeded: s/here this week/here next week/
Succeeded: s/mlist/mailing list/
Succeeded 1 times: s/chnages/changes/g