01:19
<MikeSmith>
Hixie, you pinged me yesterday
01:51
<heycam>
TabAtkins, do you mean like `new Something.OtherThing(…)?`?
01:52
<heycam>
TabAtkins, hasn't come up so far. but also I think you can just write a static operation and using new on it will just work.
02:31
<TabAtkins>
heycam: Yes, that's it.
02:32
<TabAtkins>
heycam: I'm not sure how it would work with a static operation. I could certainly define it to work as a constructor, but the definition of the returned type would presumably still have to be an interface, and thus would appear on the window, right?
02:32
<heycam>
TabAtkins, oh I see, so you want to have the interface object itself under a different object, not just an additional constructor
02:33
<TabAtkins>
Well, I suppose it's not important to hide the interface object.
02:33
<TabAtkins>
The point of putting it under another object is to curb verbosity, is all.
02:33
<TabAtkins>
But if the "real" interface exists and has a long name, that's irrelevant I suppose.
02:33
<heycam>
TabAtkins, because you get to use a shorter name?
02:33
<TabAtkins>
Yeah.
02:33
<heycam>
new Element.A("blah")
02:33
<TabAtkins>
Things like the constructors for the css value types in the new om api.
02:33
<heycam>
that kind of thing?
02:33
<TabAtkins>
Yup.
02:34
<heycam>
will all these constructors return an object that inherits from the outer thing on which you're putting the constructor?
02:34
<TabAtkins>
No.
02:34
<TabAtkins>
The outer thing will pretty much just be a namespace object.
02:35
<TabAtkins>
The plan is for a CSS interface, which is like navigator or something - a place to stash CSS-related things so they don't end up on window.
02:35
<heycam>
I see
02:36
<heycam>
well if this is a pattern that we want to go forward with I can certainly add something to the spec
02:36
<TabAtkins>
I guess hold off for now, until we have confirmation that we want to go that way.
02:36
<TabAtkins>
But provisional acceptance is good.
02:36
<heycam>
sure. I don't think it's a bad idea.
04:48
<Hixie>
MikeSmith: was going to ask if you had input on the alt thread on whatwg, from an implementor's perspective, and if you knew what hsivonen's opinion was
05:16
<MikeSmith>
Hixie: haven't caught up on replies on that thread yet
05:16
<MikeSmith>
will read it and see
05:16
<MikeSmith>
I think hsivonen is supportive of adding that attribute
05:16
<MikeSmith>
but we've been having discussions about what the validator behavior for it should be
05:17
<MikeSmith>
e.g., about whether it should cause a once-per-document warning
05:18
<MikeSmith>
Henri has explained why that would not be any better than emitting an error each time, so I'm trying to figure out some other creative way to deal with it
05:25
<MikeSmith>
anybody know who's running the OAuth 2.0 show at IETF?
05:53
<MikeSmith>
so Hixie I read Jukka's message on the alt thread on whatwg and would be happy to reply to any part of that you think I should
05:53
<MikeSmith>
I don't have anything specific to say in reply to it myself at this poitn
05:53
<MikeSmith>
*point
05:54
<MikeSmith>
I think the reply to him from Nils on that one particular question he asked says pretty much what I would have said
06:08
<MikeSmith>
Hixie: replied on the list
06:08
<MikeSmith>
if you want me to say more than what I said in that reply, just let me know
06:09
<MikeSmith>
I expect I can get something implemented in my validator.nu workspace for this over the weekend
06:58
<Hixie>
MikeSmith: awesome, thanks. reply looks perfect.
06:58
<MikeSmith>
¥OK
06:59
<Hixie>
will look more closely next week after doing the response images thread
06:59
<Hixie>
(about 25% through that)
06:59
<Hixie>
nn
14:07
<Velmont>
timeless: Yeah, that's why I said the biggest blocker :P -- But yes, I haven't investigated why it renders wrong, don't know what bug that is.
14:08
<Velmont>
timeless: And it's a "next" (aka unstable) release, so it (not being totally evil) won't autoupdate you to that ;-)