02:45
<Hixie>
ah, crap, did i forget to cc the reporter?
02:45
<Hixie>
bummer
03:23
<nonge>
problem solved - that was fast :D thanks
03:39
<Hixie>
ok, i'm running a script that ccs the reporters on all the bugs i cloned to the whatwg product for those bugs where the original reporter isn't cc'ed
06:45
<MikeSmith>
Hixie: oh thanks for running
06:45
<MikeSmith>
thanks for running that
06:45
<MikeSmith>
I had gone through and manually added a few earlier today
06:46
<MikeSmith>
Hixie: hmm is that script done running?
06:47
<MikeSmith>
because if it had run on all of those bugs I think I would be seeing a lot more bugmail in my inbox right now
06:47
<MikeSmith>
because I'm also already Cc'ed on every one of them I think
07:08
<Hixie>
MikeSmith: most bugs weren't affected
07:08
<Hixie>
MikeSmith: (most bugs are filed with my script)
07:11
<MikeSmith>
Hixie: hmm OK
07:11
<MikeSmith>
but I still think I should have gotten more than I did
07:11
MikeSmith
checks a few bugs
07:12
<Hixie>
if you find any i missed let me know
08:26
<annevk>
matjas: it should be "Platform Design Principles" prolly
08:26
<annevk>
matjas: or just "Design Principles"
08:26
<matjas>
+∞
08:27
<annevk>
matjas: I've been thinking about putting it on the WHATWG Wiki instead and shortening some of the topics and making others more generic, but I'm not sure it's worth the effort
08:28
<annevk>
(publishing an update through the W3C HTML WG is impossible at this point)
08:29
<matjas>
more divergence, yay!
08:32
<annevk>
heh
08:34
<annevk>
we've yet to see what that actually means
08:39
<hsivonen>
sigh. still someone in the comments in The Verge longing for XHTML2
08:42
<annevk>
http://xkcd.com/386/
08:55
<david_carlisle>
Hixie: thanks for addressing the bugzilla CC issue (not so much thanks for "won't fixing" my bug, but I'll think about that:-)
09:18
<hsivonen>
the HTML WG could use chairing like this: http://lists.w3.org/Archives/Public/public-webapps/2012JulSep/0212.html
09:21
<tantek>
annevk - it's likely worth the effort (putting it on the WHATWG Wiki instead and shortening some of the topics and making others more generic)
09:21
<tantek>
happy to contribute to it as well
09:27
<Ms2ger>
"Thank you for helping the HTML5 effort!"
09:27
<Ms2ger>
Hey, I wanted to help the HTML effort instead
09:28
<odinho>
:(
09:28
<odinho>
Ms2ger: Dang! You backed the wrong horse.
09:28
Ms2ger
shoots the horse
09:28
<annevk>
Ms2ger: where does it say that?
09:28
<tantek>
Ms2ger - what's the difference to you?
09:29
<Ms2ger>
annevk, http://www.whatwg.org/specs/web-apps/current-work/reviewer.js
09:29
<Ms2ger>
tantek, HTML is a useful specification, HTML5 is the W3C's waste of time
09:30
<tantek>
Ms2ger - what's wrong with a feature frozen stabilization fork?
09:30
<Ms2ger>
The HTMLWG
09:30
<annevk>
tantek: it's not maintained?
09:30
<tantek>
annevk - yes, that's certainly an issue until they empower new editors to take bug fixes from HTML.
09:31
<annevk>
you can only maintain a fork so long, because at some point features will come to depend on new features, or will be significantly rewritten, etc.
09:32
<tantek>
Ms2ger that sounds roughly like ad hominem - HTMLWG (or the people of) being bad or not does not equate to HTML5 (or the details/ideas contained therein) being bad or not.
09:32
<annevk>
and maintaining a fork is a waste of resources I think
09:32
<tantek>
annevk - I doubt features will come to depend on new features, as that would likely break backward compat.
09:32
<annevk>
publishing a snapshot for lawyers, okay, but doing anything more is a waste of time
09:32
<tantek>
I expect most new features in HTML to be reasonably modularizable.
09:32
<annevk>
tantek: they do all the time actually
09:32
<tantek>
In many ways, HTML5 could be the CSS 2.1 of HTML.
09:32
<Ms2ger>
tantek, you miss the point
09:33
<annevk>
tantek: e.g. when CORS was added existing features were defined in a different way
09:33
<Ms2ger>
tantek, the HTMLWG is the reason the HTML5 spec is bad
09:33
<tantek>
Ms2ger - I'm willing to give the yet-to-be-named HTML5 editors a chance first.
09:33
<tantek>
annevk - "defined in a different way" != break backward compat
09:34
<annevk>
exactly
09:34
<Ms2ger>
tantek, I'm not talking about hypothetical editors, I'm talking about trolls and the chairs
09:34
<annevk>
but you cannot easily take in fixes at that point
09:34
<tantek>
in practice I expect any new post-HTML5 features to be reasonably modularizable.
09:34
<tantek>
it's not clear that there's a need for further bloating a monolithic spec
09:34
<annevk>
or when we merge CORS and fetching
09:34
<annevk>
you cannot easily take such fixes in
09:34
<tantek>
I expect to see a bunch of module specs that extend HTML5.
09:34
<annevk>
going to be a mess
09:35
<Ms2ger>
tantek, who is going to write those specs?
09:35
<tantek>
Ms2ger whoever wants the new features
09:35
<tantek>
annevk - perhaps merging CORS and rewriting fetching would be a good HTML5.1 update.
09:36
<Ms2ger>
I suspect it's more likely that Hixie is going to write them
09:36
<Ms2ger>
There aren't many competent editors around here
09:36
<Ms2ger>
Anyway, I could be doing something useful
09:37
<annevk>
tantek: it's that kind of talk why I left the CSS WG; planning about drafts instead of working on tangible things
09:37
<tantek>
annevk - it's easy enough to ignore such threads/conversations if they don't interest you.
09:38
<annevk>
well yes, I quit the HTML WG
09:38
<annevk>
I have no hope they're going to be doing anything useful
09:38
<tantek>
oh? did I miss that? since the May f2f?
09:38
<annevk>
the last meeting made that pretty clear
09:38
<annevk>
since a month ago or so
09:38
<tantek>
I remember you threatening as much - or at least seeming to give up
09:38
<tantek>
I didn't realize you'd formally quit HTMLWG
09:38
<annevk>
wrote a little bit about it here: http://annevankesteren.nl/2012/07/passion
09:39
tantek
reads
09:39
<annevk>
I got chaals to remove me from the group
09:39
<tantek>
not mikesmith?
09:40
<tantek>
or did you quit webapps too?
09:40
<annevk>
your AC representative needs to remove you
09:40
<annevk>
which was chaals at the time
09:40
<tantek>
re: "Until that point HTML was by and large developed by a committee in private meetings." - the sad thing is, originally HTML was developed on an open mailing list - I'd like to know when it got moved to private lists/meetings.
09:41
<annevk>
I have not quit WebApps, WebApps is reasonably well managed, though the meetings make me cringe every now and then
09:41
<tantek>
www-talk if I recall (where HTML was developed originally)
09:42
<Ms2ger>
Heh, meetings
09:42
<tantek>
yeah, the May f2f meetings were oddly mixed.
09:43
<tantek>
Ms2ger - do you ever go to W3C f2f meetings?
09:43
<Ms2ger>
No
09:44
<Ms2ger>
https://lists.w3.org/Archives/Member/w3c-html-wg/ maybe?
09:44
<annevk>
I go to the meetings because they're fun outside of the actual meetings, the meetings themselves are mostly useless
09:44
<tantek>
annevk - re: "I quit the HTML WG politics game as what happened the other day was inevitable. " - what specifically happened the other day?
09:46
<tantek>
Ms2ger - from that I found this: https://lists.w3.org/Archives/Member/html-erb/
09:46
<annevk>
tantek: editor split
09:46
<annevk>
I guess I can add one link
09:46
<tantek>
first message to html-erb was on 1995-12-07 https://lists.w3.org/Archives/Member/html-erb/1995OctDec/0000.html
09:47
<annevk>
to http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Jul/0119.html
09:50
<tantek>
annevk - the response from Steve Faulkner is classic
09:51
<Ms2ger>
He's been on a troll lately
09:51
<Ms2ger>
*roll
09:51
<tantek>
who will be the keeper of canonical author level conformance requirements and advice?
09:51
<tantek>
Ms2ger - Freudian LOL
09:51
<Ms2ger>
tantek, we :)
10:54
<MikeSmith>
heh so the Register has chimed in
10:55
<MikeSmith>
"Judean People's Front and People's Front of Judea go their own way"
10:55
<MikeSmith>
http://www.theregister.co.uk/2012/07/23/html_5_fork/
10:56
<MikeSmith>
hmm sadly that title is the only funny part
10:57
<MikeSmith>
and that title was probably written by an editor and not by the author of the article
10:57
<odinho>
Hehe.
11:21
<tantek>
I can't imagine that many people would care on the level the Register is trying to explain - it must look like so much inside-baseball to the typical reader, even web developer.
11:22
<MikeSmith>
tantek: indeed
11:23
<MikeSmith>
and people don't read the Register to get information anyway
11:23
<tantek>
MikeSmith - I suspect we'll manage to make things work somehow regardless.
11:23
<MikeSmith>
tantek: yeah I'm sure we will
11:24
<MikeSmith>
and no matter what anybody thinks of the HTML WG for better or worse it's going to remain relevant
11:25
<Ms2ger>
*irrelevant
11:25
<MikeSmith>
cute
11:25
<MikeSmith>
but specifically in the case of document-conformance the W3C HTML5 spec going to be authoritative
11:25
<tantek>
MikeSmith - sure, I just hope those of us contributing to it can continue to do so productively.
11:25
<tantek>
(re: HTMLWG)
11:26
<Stevef_>
Ms2ger: using the troll word is a lazy way of trying to discredit a person
11:26
<Ms2ger>
Like your tweets, then :)
11:27
<MikeSmith>
tantek: yeah I do do but there are some people who have proven they often can't participate productively and who should have been permanently banned from public-html discussions long ago
11:27
tantek
gets a drink, sits back, props up his feet to watch the Msger / Stevef_ show.
11:27
<tantek>
*Ms2ger
11:28
<Ms2ger>
Have fun on your own, I've got work to do
11:28
<tantek>
*except when snarkily correcting MikeSmith ;)
11:30
<hsivonen>
it remains to be seen if Web devs end up caring about authoritative HTML5 conformance, authoritative HTML LS conformance, both or neither
11:31
<Stevef_>
I have no interest in a protracted discussion about it, just gotta call out BS when i see it
11:31
<hsivonen>
at some point, I will need to figure out a way to transition the functionality of html5.validator.nu to validator.nu and validator.nu to custom.validator.nu or complex.validator.nu
11:31
<hsivonen>
dunno how to do that without breaking the Web Service end points
11:32
<MikeSmith>
hsivonen: btw I'm wondering if maybe it's time to remove the HTML4 and XHTML1 presets from the validator UI
11:32
<MikeSmith>
I think very few people are actually using those
11:32
<hsivonen>
MikeSmith: only on Validator.nu or also on the W3C validator?
11:32
<MikeSmith>
both
11:33
<hsivonen>
MikeSmith: what about the legacy DTD stuff on the W3C validator?
11:33
<MikeSmith>
I think Jukka is the only one who has ever commented on those, and the only reason he has was to complain that the HTML4 checker doesn't follow SGML rules
11:34
<MikeSmith>
hsivonen: what about it? you mean the HTML4 facet of the W3C validator, or the 3.2 and older?
11:34
<hsivonen>
it does seem that it happens that people who think validator.nu (as opposed to html5.validator.nu) only does HTML5 validation accidentally end up doing HTML4 validation with the defaults
11:35
<MikeSmith>
ah
11:35
<hsivonen>
MikeSmith: do you mean keeping DTD-based anything around in the W3C Validator?
11:36
<MikeSmith>
hsivonen: no I don't want to keep anything DTD based around anywhere except in the legacy W3C validator
11:36
<hsivonen>
MikeSmith: but the legacy validator is staying around at the W3C?
11:36
<MikeSmith>
yeah
11:36
<MikeSmith>
b
11:37
<MikeSmith>
but eventually the Nu Markup Validator is going to be at http://validator.w3.org/
11:37
<hsivonen>
I see
11:37
<MikeSmith>
and the legacy validator will be moved to http://validator.w3.org/html4 or somewhere
11:38
<hsivonen>
but yeah, I'm not particularly enthusiastic about keeping the HTML4 schema around
11:38
<MikeSmith>
hsivonen: so I think that along with removing the HTML4 and XHTML1 presets, if we did that, we should change the doctype checking so that it doesn't ever try to use those
11:39
<MikeSmith>
basically, make the behavior like html5.validator.nu
11:39
<hsivonen>
MikeSmith: yeah
11:39
<MikeSmith>
and maybe include a message to say that if you want to do HTML4 or XHTML1 message, use the existing W3C DTD-based validator
11:39
<hsivonen>
on the Web Service side, there's the issue that validator.nu now accepts custom schemas but html5.validator.nu doesn't
11:40
<MikeSmith>
*HTML4 or XHTML1 checking
11:40
<MikeSmith>
hsivonen: I see
11:40
<hsivonen>
MikeSmith: I don't really want to encourage anyone to use DTD-based validation
11:40
<MikeSmith>
OK
11:41
<MikeSmith>
I don't want to encourage people to use XHTML1 or HTML4 doctypes
11:41
<annevk>
hsivonen: redirect html5.validator.nu to validator.nu, only expose the complex stuff as a service, and be done with it?
11:41
<hsivonen>
annevk: the thing is that the complex stuff has worse perf characteristics
11:42
<hsivonen>
annevk: so it's not nice to put the custom schema-enabled Web Service and the common-case GUI on the same host
11:42
<annevk>
does it have noticeable impact?
11:43
<annevk>
how many API consumers are there?
11:43
<hsivonen>
though maybe that could be addressed by 1) discontinuing Schematron support or 2) moving the whole thing to Heroku and hoping that superior infrastructure takes care of the user-visible issues
11:43
<hsivonen>
annevk: occasionally, validator.nu gets DoS'ed
11:43
<hsivonen>
annevk: non-malicuously AFAICT
11:44
<hsivonen>
annevk: I don't have proper metrics
11:44
<annevk>
you could ask the consumers to move to a different API during some time period
11:44
<hsivonen>
(I should probably get some)
11:44
<annevk>
and then disable it
11:44
<hsivonen>
annevk: yeah, except I have no communication channel with them
11:44
<annevk>
hsivonen: sounds like you should set one up :)
11:45
<hsivonen>
moving the validator to Heroku would be technically sensible
11:46
<hsivonen>
but it scares me to host anything in the United States
11:46
<hsivonen>
hosting it in France is legally reckless enough
11:47
<annevk>
are you scared the host will turn over data of validator users?
11:48
<hsivonen>
annevk: no, I'm scared about DMCA compliance and crazy liabilities I don't even know about
11:49
<hsivonen>
annevk: as for the EUCD, Finland and France are in the same boat anyway
11:49
<MikeSmith>
maybe switzerland
11:50
<hsivonen>
the most scary thing about France is their anti-Nazi laws
11:51
<hsivonen>
I don't intend to be a Nazi, but I don't know how crazy French enforcement is when it comes to echoing 3rd-party content (via Show Source or quotations in error messages)
11:51
<hsivonen>
I'm hoping it's clear enough that validator.nu doesn't really host any 3rd-party content it might echo
11:54
<hsivonen>
annevk: as for handing over use data, the good thing is that the validator isn't in the business of storing users' content
11:55
<hsivonen>
annevk: but it's still annoying that pretty much every competent hosting provider in Europe eventually wants to serve U.S. customers or have a data center in the U.S. to enable their European customers serve U.S. end users faster
11:56
<annevk>
makes sense though
11:56
<hsivonen>
so all your cloud data are belong to U.S. TLA agencies even if your provider is nominally European
11:57
<annevk>
I thought some of them offer options on only storing data in Europe
11:58
<annevk>
in particular for EU governments who do not want to store data in the US
11:58
<annevk>
this global internet, local laws thing sucks though
11:59
<annevk>
I think that's annoying
11:59
<hsivonen>
annevk: I doubt that helps much with staff to incarcerate in the U.S.
12:00
<hsivonen>
at least that's the general sentiment of the accidentally honest Microsoft rep
12:01
<annevk>
ooh
12:08
<hsivonen>
annevk: for [citation needed]: http://www.zdnet.com/blog/igeneration/microsoft-admits-patriot-act-can-access-eu-based-cloud-data/11225
12:08
<ruby_on_tails>
i am prepending a span in the new LIs created in a contenteditable OL using jquery, this causes issues in FF. on hitting the return key in FF, a new LI is created but the caret does not move there, you have to hit return twice to get down to the new LI created, any idea how to solve this ?
12:09
<matjas>
lolol: http://www.theregister.co.uk/2012/07/23/html_5_fork/ — “Judean People's Front and People's Front of Judea go their own way”
12:10
<ruby_on_tails>
sorry got dc
12:11
<ruby_on_tails>
basically the caret doesnt move down due to the span
12:16
<AryehGregor>
hsivonen, FWIW, I believe Wikipedia stores all its data in the U.S. specifically because U.S. laws are most favorable to its interests. The U.S. has very defendant-friendly defamation law, and strong free speech guarantees in general, relative to a typical European country. German courts have ordered wikipedia.de taken down at least once, and there have been credible complaints of infringement of British copyright law, but in both cases Ame
12:16
<AryehGregor>
rican law was clearly on Wikipedia's side.
12:16
<AryehGregor>
America is a bad place to host U.S. classified documents, gambling sites, and maybe a couple other things (some types of pornography?), but overall it tends to be safer than most European countries.
12:16
<AryehGregor>
I think the Netherlands is good too.
12:17
<AryehGregor>
That's with respect to prosecution, though. If you care about the government snooping on your data, the story might be different.
12:18
<AryehGregor>
Although I suspect most of Europe is relatively cooperative with U.S. investigations.
12:19
<hsivonen>
AryehGregor: indeed, hosting in the U.S. makes sense for Wikipedia
12:20
<hsivonen>
AryehGregor: and if you start with the assumption that U.S. TLA agencies can get your non-public data anyway, then hosting non-public data in the U.S. makes sense, because at least you are defending against snooping by *other* governments
12:21
<AryehGregor>
Right.
12:21
<AryehGregor>
For a typical site provider, the U.S. seems to be among the most legally friendly places to host your site, because it has very strong free speech guarantees in practice relative to most other countries.
12:21
<AryehGregor>
However much some people might complain.
12:22
<hsivonen>
AryehGregor: but that the U.S. makes sense for an org the size of Wikimedia Foundation doesn't mean it makes sense for someone smaller
12:22
<AryehGregor>
Well, it's true that legal protections on paper aren't very useful if you can't afford to be sued.
12:22
<hsivonen>
AryehGregor: because even if you have the law on your side, legal stuff in the U.S. is expensive
12:22
<AryehGregor>
Is it a lot cheaper other places?
12:22
<hsivonen>
AryehGregor: yes
12:22
<AryehGregor>
How so?
12:23
<hsivonen>
AryehGregor: well, insurance companies would know, and it's easier&cheaper to get legal insurance for any territory except the U.S. and Canada
12:24
<AryehGregor>
But realistically, you still strongly want to avoid a court case anyway, right? You don't want to have to pay for insurance to start with, and insurance doesn't prevent countless hours of lost time.
12:25
<AryehGregor>
(I think part of what makes some countries much better for lawsuit insurance is the loser-pays rule some countries have for civil cases.)
12:25
<hsivonen>
AryehGregor: that's a big deal
12:25
<AryehGregor>
That doesn't really help a typical individual much, because they have to pay up-front in case they do lose, but if you're insured it probably makes a huge difference.
12:25
<hsivonen>
AryehGregor: another rule that makes civil cases saner is a ban on punitive damages
12:26
<hsivonen>
AryehGregor: i.e. a principle that the damages awarded to the plaintiff can never exceed the suffered harm
12:26
<AryehGregor>
Punitive damages aren't usually a huge deal in American tort cases, I don't think. I think unreasonably high pay-outs for intangible harm like emotional damage are more of a problem, at least from anecdotal observation.
12:26
<hsivonen>
(in if you want to punish the defendant more, make it a criminal case and pay the fines to the state)
12:28
<hsivonen>
AryehGregor: anyway, hosting in the U.S. makes a great deal of sense if you have a U.S.-based legal department anyway
12:28
<AryehGregor>
Yes, that's certainly true.
12:28
<hsivonen>
AryehGregor: otherwise, not so much
12:28
<AryehGregor>
The best way to minimize legal liability in any event is to be in as few countries as possible.
12:31
<hsivonen>
(aside: I've occasionally wondered how Google dares to have a data center in Finland--a country that doesn't have proper judicial review of search warrants.)
12:32
<hsivonen>
(maybe they count on the officers who could perform a search being sufficiently afraid of Google's ability to pursue after-the-fact remedies)
12:34
<hsivonen>
(if the police in the U.S. had the kind of search powers the Finnish police has on paper, there'd be no end to abuse of police power in the U.S.)
12:45
<AryehGregor>
Fortunately, U.S. police search powers are very effectively curtailed by the exclusionary rule, since that means that it's generally pointless for them to perform illegal searches.
13:38
<AryehGregor>
Ms2ger, why do we even want contains()'s param to be "Node?" rather than just "Node"?
13:38
<Ms2ger>
Because smaug: https://www.w3.org/Bugs/Public/show_bug.cgi?id=14017
13:39
<Ms2ger>
Hmm, Opera doesn't support Node.contains?
13:52
<ruby_on_tails>
so, document.execCommand("insertBrOnReturn", false, "true"); throws error on opera, what to do ?
13:53
<odinho>
Ms2ger: >>> document.querySelector("div").contains(document.querySelector('span'))
13:53
<odinho>
true
13:54
<odinho>
Ms2ger: Seems to here.
13:55
<Ms2ger>
Huh
13:55
<Ms2ger>
Aha
13:55
<Ms2ger>
Not on Node, it doesn't
13:55
<Ms2ger>
w(document.contains(document.querySelector('span')))
13:55
<Ms2ger>
error: Uncaught exception: TypeError: 'document.contains' is not a function on line 3
14:04
<annevk>
IE only had it on Element I think?
14:04
<Ms2ger>
Possible
14:05
<annevk>
https://plus.google.com/111991826926222544385/posts/QdGfrgtP6Eg MikeSmith is blogging again, teehee
14:07
<hsivonen>
annevk: I guess being called for trolling like that means it's time for introspection. (Though see the comment I left there.)
14:08
<annevk>
I'm not sure why you were trolling
14:09
<hsivonen>
annevk: well, I wasn't in the sense that I didn't intend to troll; I just think that the main value of snapshots is the PP, which admittedly was probably too much inside baseball for a tweet
14:10
<annevk>
I have not seen any other value demonstrated
14:10
<annevk>
or explained
14:10
<annevk>
so unless I missed a memo somewhere...
14:11
<Ms2ger>
"Microsoft"
14:11
<hsivonen>
interesting that Steve's FUD spree on Twitter didn't get called out, though
14:11
<hsivonen>
Ms2ger: isn't that part of the "PP" package?
14:12
<Ms2ger>
Well, maybe if you incorrectly assume they send useful feedback...
14:13
<hsivonen>
Ms2ger: I meant in the sense that having stuff under the PP makes it more likely that Microsoft implements
14:13
<Ms2ger>
Shrug
14:14
<Ms2ger>
I tend to assume that they'll implement eventually
14:14
<odinho>
When sitecompat problems grow?
14:14
<hsivonen>
annevk: your "I'm not sure..." sentence above is a bit ambiguous. Did it seem to you like I was trolling?
14:14
<Ms2ger>
Yeah
14:14
<annevk>
hsivonen: sorry, why you were considered to be trolling
14:15
<annevk>
odinho: or when developer complaints grow
14:15
<odinho>
annevk: Ofc, that's a good place too.
14:15
<karlcow>
annevk> I have not seen any other value demonstrated
14:15
<karlcow>
I guess we all have a values system which makes it very difficult to be able to see. It's why cultures exist.
14:17
<annevk>
fortunately we have you
14:18
<Ms2ger>
Heh
14:19
<karlcow>
annevk: unfortunately I have me
14:20
<Ms2ger>
I wonder by which definition part of community moving off and creating their own version without apparent coordination is a "branch" rather than a "fork"
14:34
<Stevef_>
hsivonen: if there is any of what you call FUD here: http://www.paciellogroup.com/blog/2012/07/w3c-html-and-whatwg-html-split-affect-on-html-accessibility/ please feel free to disabuse me
14:36
<annevk>
Stevef_: e.g. the bit about AT APIs is nonsense, we did show interest in that and discussed at some F2F
14:37
<annevk>
Stevef_: it was then decided it was best to work on that separately for now
14:38
<bencc1>
is it possible that a websocket will fail after successful connection?
14:38
<bencc1>
or if a connection is established I can assume that websocket works for that user?
14:38
<annevk>
bencc1: sure, if the network drops
14:38
<Stevef_>
annevk: not nonsense at all I was told by hixie that the document was not needed and not worthwhile, that browsers had been implementing it fine already (demonstratably nonsense)
14:38
<bencc1>
annevk: I mean stuff like proxy issues
14:39
<bencc1>
annevk: if websocket fails I'm using long-polling. my question is if ws.onconnect is being called, I can assume that long-polling is not needed
14:39
<annevk>
Stevef_: does not compute with "never shown interest" though
14:40
<Stevef_>
annevk: mincing words me thinks
14:40
<annevk>
Stevef_: "never shown interest" is FUD
14:40
<annevk>
Stevef_: you can call it mincing words, whatever
14:43
<Stevef_>
annevk: apart from that?
14:43
<Stevef_>
will mark it as FUD
14:44
<annevk>
dunno
14:45
<annevk>
that you managed to get some of your pet issues through the HTML WG and not through the WHATWG does not mean the WHATWG does not care about accessibility
14:45
<annevk>
but you obviously feel otherwise about that
14:45
<annevk>
and I don't really care
14:46
<annevk>
mostly curious what this looks like a year or five years from now
14:46
<Stevef_>
annvk: i have the same antipathy towards your set of beliefs
14:47
<karlcow>
Q.E.D.
14:48
<Stevef_>
annevk: but I have decided to engage with the WHATWG never the less
14:48
<annevk>
my beliefs?
14:50
<karlcow>
☮ ♥ ✌ please
14:51
<Stevef_>
karlcow: what's so funny about peace love and understanding :-)
14:52
<karlcow>
Stevef_: nothing funny. Just need more of it.
14:52
<Stevef_>
karlcow: its a song reference if you didn't know
14:53
<karlcow>
I didn't know
14:54
karlcow
DDGed it and found Elvis Costello
14:56
<Stevef_>
Karlocw: originally by nick lowe http://en.wikipedia.org/wiki/%28What%27s_So_Funny_%27Bout%29_Peace,_Love,_and_Understanding
15:00
<Stevef_>
annevk: if you can provide anything concrete on it not being FUD, then I am happy to retract and apologise for FUD
15:05
<Ms2ger>
odinho, could you file contains-on-node or should I?
15:07
<AryehGregor>
Stevef_, I think the discussions between you and various WHATWG participants are often characterized by substantive points made on both sides that are obscured by inflammatory language, which encourages each side to dismiss the other. So for instance, in this discussion just now, both you and annevk referred to things the other one said as "nonsense". I would comment on the actual blog post you linked to, but it's not loading for me.
15:12
<Stevef_>
ArehGregor: i have attempted to be civil for the short while i have been joining the whatwg IRC, but i find that annevk in particular appears to only thinly veil or not veil at all his lack of respect for me, in such cases I find it difficult not to respond in kind, but will endeavour not to jump at the bait in future.
15:15
<odinho>
Ms2ger: Are you asking me to do it? If so, I can do it for you, but I was doing something else right now :P
15:15
<Ms2ger>
I can do it, just don't want to duplicate work
15:16
<Ms2ger>
And I don't have access to your BTS to check myself (wink nudge)
15:18
<odinho>
Ms2ger: Hmm. Yea, can't seem to find a bug like that. I'll make it. Do you have an URL handy by chance?
15:18
<odinho>
(to the place in the spec that is)
15:19
<Ms2ger>
http://w3c-test.org/webapps/DOMCore/tests/submissions/Ms2ger/Node-contains.xml / http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#dom-node-contains
15:19
<Ms2ger>
Thanks
15:19
<odinho>
Ms2ger: Cool, love it when I can just put a TC directly in without doing it :P
15:19
<Ms2ger>
So do I :)
15:21
Ms2ger
should work some more on those workers tests
15:25
<annevk>
Stevef_: All I am saying is that spreading the message that WHATWG does not care about accessibility is spreading a lie. If we did not care about it, there would be no discussion about it, longdesc would not have been removed, etc.
15:26
<annevk>
Stevef_: There is disagreement about what the right way forward is, but no disagreement about whether accessibility is a field worthy of attention
15:32
<Stevef_>
annevk: i don't believe I have said that WHATWG does not care about accessibility, what I have said is that in regards to accessibility API work, it is being done at the W3C and I or my fellow editors never got anythting positive on the work or interest shown in the work from the WHATWG side. In regards to the other aspects in the example i cited I did a put a lot of effort testing ,...
15:32
<Stevef_>
...researching, speaking with users on the effects of the accessibility of title attribute as implemented in the real world and based my arguments on that, with little or no interest in arguing the issue on its merits, the work was largely ignored or dismissed by the WHATWG. What I do belive is that accessibility issues get more consideration at the W3C, but that is something we disagree upon
15:57
<annevk>
Stevef_: I think Hixie's main point about html-api-map is that it can use some cleanup
15:58
<annevk>
Stevef_: e.g. HTML in its section WAI-ARIA already defines the default roles
15:59
<annevk>
Stevef_: so that information should not be normative in html-api-map because it's redundant
15:59
<annevk>
Stevef_: but how e.g. "link role" maps to system APIs is new information and should be normative
16:00
<annevk>
Stevef_: that's probably why Hixie largely dismissed the document
16:00
<AryehGregor>
Stevef_, I haven't followed these discussions closely, but I really think the mutual attitude of dismissiveness comes more from the way the discussion is phrased than its substantive content. The WHATWG people don't understand or agree with some of the things you're saying, this makes you get frustrated because you feel you can't make them listen to you, you lash out, they get annoyed at you and stop listening. I don't think the problem is
16:00
<AryehGregor>
that anyone doesn't care about your work; it's some combination of a) you aren't phrasing it in a way that makes sense to them, and b) they have different assumptions or priorities.
16:01
<Stevef_>
annevk: its a work in progress and if anybody bothered to look at it lately its got a lot better, none of the formation is normative currently, the ARIA stuff is there for informative purposes, some of the ARIA mappings are differnet from what is currently in HTML as they are implemnted differently
16:01
<annevk>
Stevef_: I was looking at it just now
16:01
<annevk>
maybe I got the wrong version though?
16:01
<AryehGregor>
And then both sides find it easier to just ignore the others and do their own thing, which mostly means the WHATWG people win because implementers mostly listen to them.
16:01
AryehGregor
shrugs
16:04
<Stevef_>
areygregor: not true of acc implementation stuff as there is little or no guidance from the WHATWG side, I work directly with the implementers from mozilla, webkit/chrome and we have cahnnesl into IE as one of the editors is from microsoft
16:04
<AryehGregor>
Okay.
16:04
<AryehGregor>
As I say, I haven't followed the discussions closely.
16:04
<Stevef_>
i spend a reasonable amount of my time filing bugs on the various browsers
16:05
<Stevef_>
annevk: as you know the editors draft is the only worthwhile copy http://dvcs.w3.org/hg/html-api-map/raw-file/tip/Overview.html
16:06
<annevk>
Stevef_: right so I was looking at that
16:06
<annevk>
got to go now
16:07
<Stevef_>
areyhgregor: thats because llittle of the dicsussion occurs on whatwg irc
16:07
<dglazkov>
good morning, Whatwg!
16:09
<AryehGregor>
Stevef_, I've seen the threads, I just don't read them because they don't relate to my work. I mostly do editor-related stuff for Mozilla right now, and presumably that has a11y implications somehow, but no one told me about them, so I guess someone else must be handling them.
16:10
<AryehGregor>
Also, Chatzilla should support nick autocomplete, namely typing the first few letters of a nickname and then hitting Tab to complete it. That way, you'll spell it correctly, and my IRC client will notify me that someone said my name.
16:11
<Stevef_>
aryehgregor: irc://irc.mozilla.org/accessibility is where the mozilla acc implementation work/dicussion occurs
16:11
<AryehGregor>
I'm not surprised. I also think the accessible/ dir in the source tree has something to do with it.
16:12
<Ms2ger>
I can verify that :)
16:12
<Stevef_>
aryehgregor: yes that is a mine of info as to how acc is implemented
16:19
beowulf
wanders into whatwg and finds the same basic discussion happening a year or so later
16:20
<AryehGregor>
beowulf, only one year?
16:20
<beowulf>
AryehGregor: i'm not sure, it might be two
16:38
<tabatkins_>
hober: You still planning to do something about masks in CSS?
17:03
<Hixie>
i guess today I get to look into speccing this autofill feature
17:13
<Hixie>
tantek: is the list of tokens here ok with you? i remember that you had concerns about it not matching vCard or something: http://wiki.whatwg.org/wiki/Autocompletetype#4.10.7.3.1.1_Specifying_field_data_type_hints
17:17
<tantek>
Hixie - will take a look in a bit
17:20
<annevk>
Hixie: fwiw, that list is mostly for marking up existing forms on the web including those that are not taking advantage of new features
17:21
<Hixie>
if they use this feature, they're taking advantage of new features :-P
17:22
<annevk>
Hixie: sure, what I mean is that splitting a birthday in three separate fields is not necessarily something we want to support
17:22
<Hixie>
how woudl you support the existing forms if you don't?
17:23
<Hixie>
tantek: based on the thread it looks like the list is intended to match hCard, though some people asked if it could instead match schema.org
17:24
<annevk>
Hixie: I guess the question is whether we should cater to pre-WF2-forms
17:24
<Hixie>
ah right
17:24
<Hixie>
if it's easy to do so, it doesn't seem harmful
17:24
<Hixie>
but i don't have a strong opinion one way or the other
17:25
<Hixie>
does IE do the new types yet?
17:25
<tantek>
annevk - some (many?) avoid WF2 forms controls because of the crappy UI, e.g. input type=date
17:25
<tantek>
so it's probably better to not depend on them, but do take advantage of them if they're in use
17:25
<annevk>
not sure, and I guess it makes sense to support legacy forms too, it just seems kind of odd from a conformance perspective
17:26
<tantek>
depends on if you want the feature more or less adopted, more or less practical for authors vs. purity (per design principles)
19:10
<david_carlisle>
MikeSmith: (and hsivonen) I'd agree with losing the doctype based schemas in the v.nu interfaces (at least the web facing ones, local installations might be different) although "<MikeSmith> I don't want to encourage people to use XHTML1 or HTML4 doctypes " I'd agree so it's unfortunate that the bug asking to have a doctype for xhtml+xml that wasn't one of those (and works) just got...
19:10
<david_carlisle>
...wontfixed :(
19:27
<Hixie>
david_carlisle: if there are doctypes that work and aren't in the spec, please reopen the bug listing them. I said in the bug already that if there are cases that already work it's a no-brainer to add them. It's the ones that don't already work that I wontfixed.
19:44
<Yuhong>
Personally, I use the HTML4 Transitional option mostly when I want to suppress the deprecation warnings/errors.
20:00
<Hixie>
Yuhong: why would you want to use deprecated stuff?
20:01
<Yuhong>
To validate other pages, not my own.
20:01
<Hixie>
why do you validate other pages?
20:03
<Yuhong>
For fun, mostly, but sometimes there are reasons to do so.
20:03
<Yuhong>
What is wrong with validating other pages?
20:07
<tantek>
Hixie it is good to see that the autocomplete stuff has seen some iteration to better match hCard.
20:07
<tantek>
There's actually been a lot of prior art/research on the subject here: http://microformats.org/wiki/hcard-input
20:08
<tantek>
(nothing conclusive, but the research may be useful to help inform decisions for the autocomplete stuff)
20:09
<tantek>
(aside: the suggestion to use schema instead of vCard had to either be a joke/troll, or naïvete - surprised they didn't also try using Google Base, Google Data, FOAF, or if you really want to go back, MS Passport contact schemas for that matter)
20:16
<tantek>
Hixie - shall I go ahead and edit the wiki page directly with suggested fixes and simplifications?
20:18
<TabAtkins>
Hixie: What do you think about adding a toString() to Path that produces an SVG path string?
20:18
<tantek>
ok I'm going to edit first, ask (more) questions later
20:18
<TabAtkins>
Hixie: SVGWG would like to integrate with Path as possible, and it would be cool to be able to assign one directly to a <path> or whatever.
20:20
<Yuhong>
tantek: On modularization, remember XHTML 1.1
20:20
<Yuhong>
tantek: On modularization, remember XHTML 1.1?
20:21
<tantek>
Yuhong - yes, I was in the HTML working group at the time and very much remember the problems.
20:21
<tantek>
Hence why I'm suggesting the CSS model of modularization rather than the XHTML model.
20:21
<tantek>
XHTML modularization was both a) a disaster, and b) nearly useless
20:21
<Yuhong>
Which reminds me that if I was there in 2004, I would suggest proposing a new version of XHTML that IS backward compatible with XHTML1.
20:22
<Yuhong>
Before resorting to starting the WHATWG.
20:22
<tantek>
Yuhong - see my blog posts about it from that era, e.g. http://tantek.com/log/2003/01.html#L20030114t1345
20:23
<tantek>
I was pushing for backward compatible work on XHTML1/HTML4 back in 2003 - *in* the HTML Working Group itself. Unfortunately I was the only one advocating this focus (in the WG).
20:23
<tantek>
so yes, I've witnessed firsthand both a modularization failure (XHTML) and a modularization success (CSS2.1/3)
20:23
<Yuhong>
And unfortunately what got submitted by Mozilla/Opera was not that.
20:24
<tantek>
thus given that data, I'm advocating CSS-style modularization, where we agree to modularize on top of a monolithic spec, rather than to replace it.
20:24
<Yuhong>
Back in 2004, of course.
20:24
<tantek>
thus as we have CSS 2.1, we now have HTML5
20:24
<tantek>
and we have CSS3 modules that build on top of CSS2.1, we can have HTML modules that build on top of HTML5.
20:27
<Yuhong>
Sorry, did not realize the difference between CSS and XHTML.
20:27
<Yuhong>
modularization.
20:27
<tantek>
there's more than one way to modularize
20:27
<Yuhong>
Agreed.
20:28
<tantek>
so it's easy enough to pick one that is mostly working, and improve on it, rather than the one that was a nearly a completely failure
20:29
<Yuhong>
Of course.
20:29
<tantek>
Yuhong - you'll notice that WHATWG does tend to leave CSS stuff to the W3C CSS WG - because it tends to do a good enough practical job that it's not worth reinventing.
20:29
<tantek>
it's not perfect, but no standards org/community is.
20:35
<Hixie>
yutak: nothing wrong, just curious :-)
20:35
<Hixie>
tantek: cool, thanks for the link
20:35
<tantek>
regarding the autocomplete stuff, it's funny that it fails to mention the prior art by Microsoft - the VCARD_NAME field: http://msdn.microsoft.com/en-us/library/ms533032%28VS.85%29.aspx
20:35
<Hixie>
tantek: ideally post them to the list with rationale
20:35
<tantek>
(research that's in the microformats wiki on this subject)
20:35
<Hixie>
tantek: i'm probably adding them to the spec today or tomorrow, but they can always change after
20:36
<tantek>
judging by the screenshots, I'd estimate MS has supported that since IE6
20:36
<tantek>
Hixie - do you care at all about IE's implementation?
20:36
<tantek>
or should we design a brand new feature here?
20:36
<Hixie>
tabatkins: seems not unreasonable, if svg paths can describe everything canvas paths can do... i guess someone would have to come up with an algorithm
20:36
<tantek>
(I'm not saying there's an obvious right answer - I'm just looking for your general thinking on the matter)
20:37
<Hixie>
tantek: i care about all relevant data
20:37
<Hixie>
there's been a bunch of attempts at addressing this
20:37
<Hixie>
post mortems for those are what i'd really love to see
20:37
<tantek>
"shipping in IE for many years" is a bit stronger than "attempt" ;)
20:38
<tantek>
as I think you said when this topic first came up, it's not clear how many web developers care about this at all
20:38
<Hixie>
netscape shipped support for ECM for years too
20:39
<Hixie>
there's lots of shipped failures in this field :-(
20:39
<tantek>
I think IE might still support this
20:39
<tantek>
longest continuous (still) shipping implementation likely counts for more than abandoned attempts
20:40
<tantek>
maybe I'll ask on Twitter if anyone has heard of or uses it, or if IE even supports it still
20:40
<tantek>
do you have a link to the "netscape shipped support for ECM" documentation so I can add it to my research?
20:40
<Hixie>
i do not
20:42
<tantek>
ok I'll mention it as anecdotal then - let me know if you do find something
20:42
<tantek>
do you have a rough year range of when netscape shipped support for ECM?
20:43
<Hixie>
6.x range
20:43
<Hixie>
iirc
20:43
<Hixie>
man, searching the net for vcard_name is unfulfilling
20:44
<Velmont>
Yeah, if only someone here worked at a company knowing lots about search or something... :/
20:45
<Velmont>
:P
20:46
<Hixie>
?
20:46
<Hixie>
i don't mean the results are bad
20:46
<Hixie>
i mean the results show that people don't really know what it is
20:47
<tantek>
Hixie, are you searching HTML source for it?
20:48
<Hixie>
i was looking for tutorials and examples of people using it
20:48
<tantek>
for usage stats?
20:48
<tantek>
the only tutorials I know of are on MSDN
20:48
<tantek>
for VCARD_NAME
20:48
<tantek>
like most microsoft-only extensions
20:48
<Hixie>
hence my using a search engine to look for others :-P
20:49
<tantek>
can you do a search over crawl data for VCARD_NAME in the source? I'm curious about adoption in the wild.
20:49
<tantek>
(I don't know of any actual sites that use it - would be useful to have a few examples to look at)
21:05
<tantek>
I'm going to ask folks in #microformats to take a look at our research and iterate on it as well
21:06
<tantek>
anyone interested in autocomplete / vCard auto-fill type stuff should feel free to jump into #microformats and contribute - we're a fairly friendly bunch
21:06
<tantek>
(though we do ask for the usual, use-cases, data, real world examples, and sometimes tend to summarily dismiss theoretical examples and wishful thinking)
21:20
<annevk>
tantek: fwiw, the CSS-stye modularization is starting to become a mess for at least implementors and specification precision
21:20
<annevk>
not sure I'd classify it as a success
21:22
<tantek>
annevk - if you mean "total success" no
21:22
<tantek>
however *a* success, certainly yes, based on two very easy metrics
21:22
<tantek>
1. for spec-types - modules have made it to REC
21:23
<tantek>
2. for devs - browsers have interop on several modules
21:23
<Hixie>
making to REC is hardly a measure of success :-)
21:23
<Hixie>
2 is valid though
21:23
<tantek>
Hixie - I knew you were going to say that :P
21:23
<tantek>
perhaps I should have said "for process-type" rather than "for spec-types" for 1.
21:24
<tantek>
annevk - also, I consider overall CSS modularization a success especially when *some* modules fail.
21:24
<Hixie>
the biggest problem with CSS is the vagueness of the specs
21:24
<tantek>
that means that we successfully isolated features that the market was not as interested in, and stopped them from holding back features that the market wanted
21:25
<Hixie>
but that's more or less inevitable for the modules because they all rely on a foundation that is too vague
21:25
<tantek>
Hixie - s/CSS/specs in general
21:25
<Hixie>
no, not any more
21:25
<annevk>
euhm, features that the market wanted? you mean like a better layout model?
21:25
<annevk>
I mean sure, some tiny stuff was successful, such as color
21:25
<Hixie>
tantek: XHR, HTML, Origin, CORS, lots of the new APIs -- lots of new specs these days are written very tightly and not vaguely
21:25
<annevk>
but that's not really what most people are waiting for
21:25
<tantek>
Hixie - I think "too vague" is a bit harsh - based on the 2.1 test suite and impl reports.
21:26
<Hixie>
tantek: don't get me wrong, CSS is way better than what came before the 2.1 era
21:26
<Hixie>
tantek: but it hasn't kept improving, and is now no longer close to cutting edge in terms of precision
21:26
<tantek>
and as I said, CSS modularization is not perfect at all - it's just shown a degree of success (in contrast to other modularization attempts which have shown nearly total faliure)
21:27
<Hixie>
it's aroudn the same level of precision as SVG specs, which is to say, way better than the status quo around 2000, but not at the level we expect in 2010
21:27
<tantek>
Hixie - we just need to get TabAtkins to stop renaming Flexbox properties ;)
21:27
<tantek>
Hixie - fair enough, the bar for writing specs has been raised.
21:27
<Hixie>
what things are named isn't really a matter of precision :-)
21:28
<tantek>
even though we know how to write better specs, I would offer that lack of spec precision isn't stopping implementers of the CSS modules
21:28
<tantek>
so yes, the specs *can* be better written, however, they are doing a 'good enough' job to help the platform move forward
21:29
<tantek>
could we do an even better job? of course. do we have enough editors to do so? no. so we settle for incremental improvements.
21:30
<Hixie>
clearly lack of precision has never stopped implemntors
21:31
<Hixie>
witness HTML in the 90s
21:31
<Hixie>
or CSS in the 90s
21:31
<Hixie>
or HTTP
21:31
<Hixie>
or...
21:31
<Hixie>
it just results in poorer interop
21:32
<tantek>
right, and as precision has improved, so has interop
21:32
<Hixie>
precision and test suites, yeah
21:32
<tantek>
thus measured by interop, CSS 2.1 (incl test suite) has done a decent job
21:33
<Hixie>
for what's tested
21:33
tantek
looking forward to seeing what gets dropped from HTML5 during CR due to lack of impl/interop per test suites.
21:33
<Hixie>
i mean, entire sections of that spec are intentionally undefined
21:33
<Hixie>
e.g. the table model
21:38
<tantek>
yeah
21:39
<tantek>
Hixie, re: autocomplete, would it be reasonable to ask you to shuffle your task items to postpone spec'ing it for a week to see what the microformats community can come up with in terms of updating the research documented at http://microformats.org/wiki/hcard-input and sub-pages?
21:39
<tantek>
have just today/tomorrow is a bit short notice.
21:45
<Hixie>
it's hardly short notice, you participated in the threads back in january :-)
21:47
<tantek>
is it reasonable to ask for a week of updating research etc.?
21:48
<tantek>
back in January (or was it December?) it seemed like a Google-only thing
21:48
<tantek>
that didn't do any previous research etc.
21:48
<tantek>
now that it's gotten some basis in previous work, it's more interesting
21:50
<Hixie>
well the spec can always change :-)
21:50
<Hixie>
i haven't done anything yet, but it's what i'm workong on now
21:51
<tantek>
sure, the spec can always change, however you've shown a pattern of less change rather than more after a first iteration "lands" in the spec
21:52
<tantek>
also - implementers seem to take your first landing of a feature as a go signal to implement - so it has a lot of weight (even if unintentionally so) - gets press articles written, etc.
21:52
<Hixie>
i do less change after something exists to change than before something exists to change? o_O
21:52
<tantek>
sometimes even flamewars ;)
21:53
<Hixie>
well chrome already implements something on these lines, so that's not a concern here :-)
21:53
<tantek>
btw - overall comment - too many feature ;)
21:53
<tantek>
features
21:53
<tantek>
in autocomplete
21:53
<Hixie>
anyway, if you want to send feedback, send feedback :-)
21:54
<Hixie>
i assure you i take it as much into account if it's sent before i've looked at a thread as if it's sent after
21:54
<tantek>
will do
21:55
<tantek>
I'd just rather take the time to send you feedback with citations rather than feedback off the top of my head.
21:56
<tantek>
also - has anyone seen isherman around? would love to be able to do some of this interactively with him (since he wrote up and iterated http://wiki.whatwg.org/wiki/Autocomplete_Types )
22:07
<Hixie>
hmm, https://src.chromium.org/viewvc/chrome/trunk/src/chrome/test/data/autofill/heuristics/input/ is apparently a list of forms synthesised from real world forms which apparently the chrome team did when they originally came up with this
22:18
<tantek>
Hixie - thanks - will add that to the hcard-input-examples page for further analysis
22:23
<tabatkins>
tantek: Slanderer. ^_^ I stopped renaming things once we entered LC, and successfully blocked an attempted rename recently.
22:28
<tantek>
TabAtkins - I get all my news from W3Cmemes.
22:29
<tabatkins>
There's your problem.
22:52
<Hixie>
tantek: why is the country in hCard country-name rather than country?
22:52
<tantek>
because that's the phrasing the vCard spec used and in the interest of minimal change I chose their literal phrasing
22:52
<tantek>
same with "postal-code"
22:52
<Hixie>
o_O
22:53
<tantek>
(modulo singularization of plural names)
22:53
<Hixie>
seems a bit silly to derive a formal term from inconsistent prose rather than making it consistent :-)
22:53
<tantek>
nah, the prose was fairly consistent across properties
22:53
<tantek>
as what happens when people copy/paste property descriptions
22:53
<Hixie>
vcard 4 uses just "country" in the BNF
22:53
<tantek>
oh does it
22:53
tantek
looks
22:55
<tantek>
alrighty then, first google result for "vcard4" is a wiki page I created. thanks google.
22:55
<Hixie>
hehe
22:55
<Hixie>
i get that kind of stuff all the time
22:55
<Hixie>
search for some obscure thing and the first hit is some e-mail i wrote asking for help on the very topic
22:56
<tantek>
Hixie - in 6.3.1 it uses "country name" http://tools.ietf.org/html/rfc6350#section-6.3.1
22:56
<Hixie>
yes but that's the prose
22:56
<Hixie>
the formal bnf uses just "country"
22:56
<tantek>
historically the prose tends to be more accurate
22:56
<tantek>
since that's what human read/review
22:56
<Hixie>
the prose here is inconsistent
22:56
<tantek>
tons of problems in the grammar have been found over time
22:56
<Hixie>
uses "country name" but not "city name", e.g.
22:57
<Hixie>
anyway, my point is that it's dumb to use "-name" on the country field and not the others, it adds nothing :-)
22:57
<tantek>
so you're arguing from a shortening/simplicity perspective
22:57
<tantek>
to make an exception
22:57
<tantek>
b
22:57
<tantek>
e
22:57
<tantek>
ca
22:57
<tantek>
us
22:57
<tantek>
e
22:57
<tantek>
(darnit)
22:58
<tantek>
because the prose is fairly consistent in 6.3.1
22:58
<tantek>
the post office box;
22:58
<tantek>
the extended address (e.g., apartment or suite number);
22:58
<tantek>
the street address;
22:58
<tantek>
the locality (e.g., city);
22:58
<tantek>
the region (e.g., state or province);
22:58
<tantek>
the postal code;
22:58
<tantek>
the country name (full name in the language specified in
22:58
<Hixie>
i'm arguing for consistency within the grammar itself, rather than with prose that is internally inconsistent
22:58
<tantek>
the {propertyname} {(optional parenthetical remark} ;
22:58
<Hixie>
s/grammar/vocabulary/
22:58
<tantek>
where is the prose internally inconsistent?
22:58
<Hixie>
it says "locality" not "locality name", and "country name" not "country"
22:59
<Hixie>
yet gramatically those (and region) should be equivalent
22:59
<tantek>
well if it used "name" as a suffix on all of them, then it would be consistent to drop it on all of them
22:59
<tantek>
I only kept -name on country because that was the only one that had it
22:59
<Hixie>
i understand what you did, i'm just saying it's silly :-)
22:59
<tantek>
it's not silly, it's just literal :)
23:00
<tantek>
so you think it's worth 5 characters to break from that?
23:00
<tantek>
or apply a new property naming rule of drop "-name" from properties from other formats?
23:00
<Hixie>
i think basing names on prose that isn't giving normative vocabulary terms is something that must be done with great care and not blindly
23:01
<tantek>
Hixie - I think a methodology that's more easily reproduceable (in the scientific sense) is better than a picky/careful one that may result in different results depending on who applies it and their opinions
23:01
<tantek>
hence a simpler methodology tends to be better
23:01
<tantek>
anyway - that was done nearly 8 years ago, so now you want to drop the 5 characters?
23:02
<Hixie>
not in hcard, no, of course not :-)
23:02
<tantek>
are you going to want to rename "org" to "organisation" [sic] like the RDF folks did?
23:02
<tantek>
or worse, "orgName
23:02
<tantek>
"
23:02
<tantek>
?
23:02
<Hixie>
hcard is shipped, there's not much one can do
23:02
<Hixie>
i'm just being an armchair critic :-)
23:03
<tantek>
I'm on the fence about it
23:03
<tantek>
over the years it has bugged me, but I haven't had sufficient reason to change it
23:03
<Hixie>
but i strongly disagree with your description of how to design a vocab. language design isn't something you should do by methodically transferring terms from a description of another vocab to the new vocab, it's more of an art form, imho
23:03
<tantek>
and IMHO that's the right way with standards, don't change something unless there is a good reason to do so
23:04
<tantek>
nah - the folks that claim "artform" are the ones that create new vocabs out of laziness
23:04
<tantek>
they don't do their home work
23:04
<tantek>
or they want to reinvent the world
23:04
<tantek>
(arrogance, ego)
23:04
<tantek>
like most of schema.org
23:04
<Hixie>
i'm not saying it should be done in the dark
23:04
<gavinc>
tantek: +1
23:04
<tantek>
but that's what happens in practice
23:05
<Hixie>
you obviously want to derive new vocabs from existing vocabs, and even more, from what is learnt from those vocabs having shipped
23:05
gavinc
goes back to creating vocabs as an art form
23:05
<tantek>
folks that claim vocab is an artform are usually doing so because they're too lazy to be scientific about it
23:05
<Hixie>
i'm just saying that you have to _apply_ those learnings when creating a new vocab, not just copy the mistakes on down, or introduce new ones without a need
23:06
<tantek>
there is some art to vocab design yes, but there is a heck of a lot more principled / constraint based design that goes into it
23:06
<gavinc>
better to copy the old mistake then mostly copy the old mistake and make a bit worse
23:06
<tantek>
Hixie - most folks that try to "fix" old "mistakes" make *worse* mistakes
23:06
<tantek>
that's the problem
23:06
<tantek>
and typically no two people fix the same mistake in the same way
23:06
<Hixie>
you're not most folks
23:06
<tantek>
so you get divergence
23:06
<Hixie>
i expect better of you :-)
23:06
<tantek>
and babble
23:06
<tantek>
hah!
23:07
<tantek>
scientists aren't supposed to assume that they're special (as an individual)
23:07
<Hixie>
language design isn't just a science
23:07
<tantek>
there are exceptions to the "fix" old "mistake" divergence pattern
23:07
<tantek>
the biggest exception I know of
23:07
<Hixie>
it's, to possibly coin a new phrase, evidence-based art
23:07
<tantek>
is that multiple independent parties have renamed "fn" to "name"
23:08
<tantek>
so given that experience of seeing that happen
23:08
<tantek>
we've done that as well in microformats2
23:08
<danbri>
how is microformats2 coming along?
23:08
<tantek>
evidence-based art - I like that
23:08
<tantek>
danbri - quite well
23:08
<danbri>
i dropped a mail to the list a while back, didn't see a reply... but i liked the general direction
23:09
<tantek>
danbri - I replied!
23:09
<tantek>
I think you were even explicitly cc'd on it
23:09
<danbri>
hmm, mail snafu then, investigating!
23:09
<tantek>
could be on either end
23:09
<tantek>
or perhaps email IS the snafu ;)
23:10
<tantek>
Hixie - have you seen other vocabularies "correct" 'country-name' to just 'country" ?
23:10
<Hixie>
tantek: vcard4 :-)
23:10
<tantek>
if you can cite examples of independently agreeing corrections, it's worthy of consideration
23:10
danbri
remembers going looking to double check it was all quiet :( searching "microformats 2" now I see an unread Tantek reply
23:10
<tantek>
hah - that's just the grammer production being shortened to fit in 70 char widths
23:10
<tantek>
grammar even
23:10
<tantek>
a-ha!
23:11
<tantek>
so you did get the reply. well at least there's that.
23:11
<Hixie>
tantek: adding "-name" is not a length problem, it would still fit
23:11
<danbri>
ah ok it was 20 days later, likely I gave up after a fortnight assuming a billion mails would have swamped it in your inbox subsequently
23:11
<tantek>
danbri - microformats2 is getting a good reception among web designers
23:11
<danbri>
thanks for replying, sorry I missed it!
23:11
<tantek>
the simpler design in particular is greatly applying
23:11
<Hixie>
tantek: i think they just did it for consistency in the grammar (though i'm just guessing) and thought the prose scanned better with some having "name" and some not
23:11
<tantek>
appealing even
23:11
<tantek>
s
23:11
<tantek>
h
23:11
<tantek>
ees
23:11
<tantek>
h
23:11
<tantek>
it may be coffee time
23:11
<tantek>
Hixie - it's been "country-name" since...
23:12
<danbri>
sleep o'clock here
23:12
<danbri>
i'll check in w/ Ben Ward re parsing
23:13
<Hixie>
on another note, is there a vcard/hcard extension for IMs?
23:13
<tantek>
danbri - feel free to ping on #microformats any time
23:13
<Hixie>
do people just do a new field per IM type?
23:13
<tantek>
Hixie - yes, impp
23:13
<danbri>
will do, cheers
23:13
<Hixie>
or is there a set of terms for "im-type" and "im-whateveryoucallthevalue"?
23:13
<tantek>
it's been incorporated into vCard4
23:13
<tantek>
nah
23:13
<tantek>
it's just a variant of 'url'
23:13
<gavinc>
nwalsh's last version of the vcard owl lines up with hCards naming too, not that too many people care
23:13
<tantek>
the "type" is still URL
23:13
<tantek>
gavinc - URL?
23:14
<Hixie>
what url do you use for, say, aim? or icq?
23:14
<tantek>
Hixie - see tantek.com for an example aim: URL :D
23:14
danbri
reminded of PortableContacts threads about OpenSocial I meant to read up on
23:14
<gavinc>
tantek: http://nwalsh.com/rdf/vCard
23:14
<Hixie>
is that a scheme that anyone actually consumes?
23:15
<tantek>
Hixie - yes - if you click it, and have e.g. Adium installed, or iChat, it will work
23:15
<Hixie>
wikipedia says people use x-aim, x-icq, x-jabber, x-msn, x-yahoo, x-twitter, x-skype-username...
23:15
<Hixie>
in vcard
23:15
<tantek>
Hixie, re: aim:, Yahoo, MSN, Skype, all documented here: http://microformats.org/wiki/hcard-examples#New_Types_of_Contact_Info
23:15
<gavinc>
It's a scheme that has been used to store hCard extracted data in SPARQL stores
23:16
<gavinc>
but rdf data tends towards foaf over vcard
23:17
<Hixie>
tantek: thanks, perfect
23:18
<gavinc>
the most intesting part being "This ontology attempts to model a subset of vCards in RDF using modern (circa 2005) RDF best practices. The subset selected is the same subset that the microformats community has adopted for use in hCard."
23:19
<tantek>
ICQ… hmm...
23:19
<tantek>
anyone use ICQ and know how to construct icq: URLs?
23:20
gavinc
learns that ICQ still exists
23:20
<zewt>
every IM ends up still existing forever, for the isolated pockets of people who use it among each other, heh
23:21
<zewt>
(was: re: irc)
23:21
<tantek>
Hixie - which wikipedia article did you find about x-aim, x-icq, x-jabber, x-msn, x-yahoo, x-twitter, x-skype-username ?
23:21
<gavinc>
Yeah, was about to say ... he says on IRC ;)
23:21
<Hixie>
tantek: http://en.wikipedia.org/wiki/VCard
23:21
<tantek>
"a number of which are in common usage." yeah right.
23:21
<tantek>
CITATION NEEDED
23:21
<tantek>
;)
23:22
<zewt>
heh
23:22
<Hixie>
my citation was wikipedia :-)
23:22
<zewt>
also it's 2012 and every messaging system still has its own isolated client; the future is pathetic
23:22
<zewt>
all I want is to be able to contact people without caring what protocol they use or what protocol I'm using, let me pick a person and send a message and have it work
23:23
<zewt>
bluh i hate bad defaults, because inevitably it means that when you change it to the right setting, it ends up being an obscure buggy code path
23:24
<tantek>
zewt, Adium works reasonably well
23:24
<tantek>
indeed, we've regressed since mailto:
23:24
<zewt>
re: changing firefox to not allow pages to open new windows/tabs, which works--except it looks like if a page opens what would normally have been a new window (which navigates into the same window under these settings), it means the page can close the tab even though it wasn't actually opened by script
23:24
<tantek>
and btw, updated: https://en.wikipedia.org/wiki/VCard#vCard_extensions
23:24
<tantek>
:P
23:24
zewt
grump gripe
23:26
<tantek>
Hixie - please tell me you're not doing format/vocab design based on wikipedia documentation of asserted vCard extensions - most of those are crap / obsolete.
23:26
<tantek>
as they say in school (journalism?), Wikipedia is not a primary source, does not itself make for a good citation. It may help you *find* primary sources / good citations, but that's about it.
23:28
<tantek>
btw just to point out how badly out of date wikipedia can be - that vCard article doesn't list vCard4 properties here: https://en.wikipedia.org/wiki/VCard#Properties
23:29
<gavinc>
tantek: http://hg.adium.im/adium/file/9f195727e81d/Source/AIURLHandlerPlugin.m#l139 is what your referring to for Adium>
23:29
<gavinc>
?
23:29
<tantek>
http://www.adium.im/
23:29
<tantek>
is what I've got installed
23:29
<tantek>
and am reasonably happy with
23:30
<gavinc>
I meant the list of schemes :P
23:30
<tantek>
it's certainly better than the default iChat
23:30
<tantek>
oh look at that
23:30
<Hixie>
tantek: i use all kinds of sources, but yeah, i do use wikipedia
23:30
<tantek>
just as long as you keep in mind how obviously flawed it can be
23:30
<tantek>
all sources are not equal
23:30
<Hixie>
oh don't worry, i've read the HTML-related pages
23:30
<tantek>
hahaha
23:32
<tantek>
gavinc that's an interesting list of schemes, but not enough documentation to actually construct working URIs with each scheme
23:32
<tantek>
which is what the hCard examples provide: http://microformats.org/wiki/hcard-examples#New_Types_of_Contact_Info
23:33
<tantek>
e.g. how do you construct an irc: URI to open a private message window to particular nickname?
23:33
<tantek>
I know this works for a channel: irc://irc.freenode.net/microformats
23:33
<tantek>
but how do you message a nickname on a network?
23:36
<tantek>
nm re: icq scheme - there isn't one. apparently we documented that too: http://microformats.org/wiki/hcard-examples#ICQ
23:41
<Hixie>
tantek: i sent an e-mail rather than updating the spec, with a proposal. so please reply to it soonish if you want to reply before i edit the spec.
23:43
<tantek>
Hixie, I'm surprised to see you suggest "x-" prefixes
23:43
<tantek>
I thought for sure you would be against those
23:43
<tantek>
even the IETF recently obsoleted the "x-" prefix practice
23:43
<Hixie>
for things that are purely experimental and don't ship?
23:43
<tantek>
the problem is things leak
23:44
<Hixie>
there's a place for x- prefixes
23:44
<Hixie>
the problem is when people use them for extensions, rather than experiments
23:58
<gavinc>
<name of person doing experiment>-blah
23:58
<gavinc>
That's unlikely to leak ;)