| 00:17 | <Xano__> | Is it usual to expirience a long delay between requestion the current position using geolocation and the excution of the callback? Sometimes I don't get so anything, even after a minute |
| 00:22 | <AryehGregor> | Xano_, maybe the user didn't give permission. |
| 00:23 | <AryehGregor> | Generally it will pop up a bar asking if the user wants to allow the location to be used. |
| 00:23 | <AryehGregor> | If not, then you won't ever get it. If so, you'll still only get it when the user gets around to clicking. |
| 00:25 | <MikeSmith> | AryehGregor: looks like dud left already |
| 00:26 | <MikeSmith> | 2 minutes after asking his question |
| 00:26 | <AryehGregor> | He left and came back. |
| 00:26 | <MikeSmith> | oh |
| 00:26 | <AryehGregor> | [101107 19:24:36] * Xano__ has quit (Read error: Connection reset by peer) |
| 00:26 | <MikeSmith> | I see he's still here |
| 00:26 | <Xano__> | AryehGregor: Yeah, my connection went dead for a moment |
| 00:26 | <Xano__> | AryehGregor: I got the permission dialog |
| 00:27 | <Xano__> | And I believe I checked the checkbox to remember my permission |
| 00:28 | <AryehGregor> | What browser? |
| 00:33 | <Xano__> | And there it happened again |
| 00:34 | Xano__ | blames his flatmates |
| 00:35 | <MikeSmith> | Xano_: why is that not just something that's implementation dependent? |
| 00:37 | <Hixie> | Xano__: it can take a long time if it's spinning up the GPS receiver, that might be it |
| 00:38 | <Xano__> | MikeSmith: I believe it *is* implementation dependent, but because I guess folks here have tested those implementations I figured I'd ask here |
| 00:38 | <Xano__> | Hixie: I'm testing in FF now |
| 00:38 | <Xano__> | Forgot to mention that, sorry |
| 00:39 | <Philip`> | Hixie: I didn't know GPS receivers used any kind of spinning gyroscope or flywheel |
| 00:40 | <AryehGregor> | Philip`, of course they do, how else would they attract the satellites' attention? |
| 00:41 | <Xano__> | AryehGregor: They don't. |
| 00:41 | <Xano__> | AryehGregor: They use their built-in towel to know where they are |
| 00:41 | <AryehGregor> | But their towel ceases to function indoors? |
| 00:42 | <AryehGregor> | All my towels are indoors. Maybe if I brought them outdoors, they'd work as location sensors too, rather than just things to dry off other objects with. |
| 00:42 | <Xano__> | AryehGregor: Not if you carry it around your head |
| 00:42 | Xano__ | is going to order a towel with "GPS receiver" on it |
| 00:46 | <MikeSmith> | revolutionary new browser whose only difference from any existing browser seems to be that it requires users to have a facebook account |
| 00:47 | <MikeSmith> | wish I could get investors to give me 10 million dollars to come up with an idea like that one |
| 00:47 | <AryehGregor> | You're supposed to come up with the idea before they give you the money, usually. |
| 00:47 | <MikeSmith> | ah |
| 00:48 | <AryehGregor> | . . . also, which browser? |
| 00:48 | <MikeSmith> | well, in that case they should be paying their 10 million to the Flock people |
| 00:48 | <MikeSmith> | because I can't see there's anything here that Flock didn't do 5 years ago or whatever |
| 00:48 | <MikeSmith> | AryehGregor: RockMelt |
| 00:49 | <MikeSmith> | something Marc Andreesen is involved with, apparently |
| 00:49 | <AryehGregor> | It's remarkable how a name can evoke such visceral contempt before I have the faintest clue what the thing is. |
| 00:50 | <MikeSmith> | heh |
| 00:50 | <MikeSmith> | yeah, Marc Andreeson seems to have that effect on people |
| 00:50 | <MikeSmith> | oh |
| 00:50 | <MikeSmith> | I guess you meant the name RockMelt |
| 00:50 | <MikeSmith> | "Netscape founder seeks to revolutionize Web surfing again" |
| 00:50 | <AryehGregor> | "Hopefully as profitably as the last time" |
| 00:51 | <MikeSmith> | "RockMelt, a new browser that lives in the cloud" |
| 00:51 | <MikeSmith> | so that's revolutionary at least |
| 00:52 | <MikeSmith> | living in the cloud |
| 00:52 | <MikeSmith> | I guess that means I don't even have to download it |
| 00:52 | <AryehGregor> | I'm pretty sure there are all sorts of mythologies where various beings live in the clouds. |
| 00:52 | <AryehGregor> | And build castles there and everything. |
| 00:52 | <Hixie> | Philip`: they use hamster wheel |
| 00:52 | <MikeSmith> | Star Trek |
| 00:52 | <AryehGregor> | Unless Heroes of Might & Magic lied to me. |
| 00:52 | <MikeSmith> | cloud people |
| 00:52 | <AryehGregor> | It says titans live in cloud castles. |
| 00:53 | <MikeSmith> | http://en.wikipedia.org/wiki/The_Cloud_Minders |
| 00:53 | karlcow | is using eagle feathers, skin of frogs, and moose horns to find my way around |
| 00:55 | <MikeSmith> | each Star Trek reference should have a special section that lists the name of whatever scantily but not too scantily clad alien babe Kirk has discretely implied sex with |
| 00:55 | <MikeSmith> | for each episode, I mean |
| 00:56 | <Philip`> | http://www.vgmaps.com/Atlas/SuperNES/ChronoTrigger-12000BC-KingdomOfZeal.png - they even had cloud castles in caveman times |
| 00:57 | <AryehGregor> | Living in the cloud seems impractical. They aren't dense enough to provide sound structural support. |
| 00:58 | <AryehGregor> | I've heard that the atmosphere of Venus can be dense enough for habitation, though. Maybe they mean the clouds of Venus. |
| 00:59 | <Philip`> | Why do you need structural support? |
| 00:59 | <AryehGregor> | http://www.universetoday.com/15570/colonizing-venus-with-floating-cities/ |
| 00:59 | <Philip`> | Just use something equivalent to a helicopter |
| 00:59 | <Philip`> | plus a clever refuelling system |
| 00:59 | <AryehGregor> | Then you're really only using the cloud for what, cover from potential enemies? Otherwise you could just as well use any old patch of sky. |
| 00:59 | <Philip`> | Or, I suppose, a hot air balloon |
| 01:00 | <Philip`> | Use the cloud for fresh water |
| 01:02 | <AryehGregor> | Hmm, I see. |
| 01:02 | <AryehGregor> | What sort of cloud would you want? I imagine more like cumulus or stratus clouds. |
| 01:02 | <AryehGregor> | What do you do if the cloud dissipates? |
| 01:04 | <Philip`> | You migrate to another cloud |
| 01:04 | <AryehGregor> | Sounds great. How much does it cost to sign up? |
| 01:05 | <MikeSmith> | http://en.wikipedia.org/wiki/Castle_in_the_Sky |
| 01:05 | <Philip`> | That's the beauty of it - you only have to pay for what you use |
| 01:05 | <MikeSmith> | Laputa |
| 01:05 | <Philip`> | MikeSmith: They totally stole that from Gulliver :-( |
| 01:06 | <MikeSmith> | most of Miyazaki films are based on ideas from somebody else |
| 01:07 | <MikeSmith> | but he just uses the idea to hang all kinds of other stuff on |
| 01:07 | <Philip`> | Better not hang too much stuff on, or it'll fall out of the sky |
| 01:07 | <MikeSmith> | he sort of just keeps making the same movie, and just changing the story a little |
| 01:08 | <MikeSmith> | dudes |
| 01:08 | <MikeSmith> | "HTML 5 is the glue for the cloud, says Ballmer2 |
| 01:08 | <AryehGregor> | That's mixing your metaphors all right. |
| 01:08 | <MikeSmith> | *Ballmer |
| 01:09 | <MikeSmith> | AryehGregor: he's clearly anticipated already the problem you described |
| 01:09 | <MikeSmith> | the solution is glue |
| 01:09 | <MikeSmith> | you just pump a bunch of glue into the cloud |
| 01:09 | <MikeSmith> | and that glue is made of pure unadulterated HTML5 |
| 01:12 | <MikeSmith> | man, Alexander Clauss is still at it with iCab |
| 01:12 | <MikeSmith> | http://pmptoday.com/apple/6379 |
| 01:12 | <MikeSmith> | "iCab Mobile is a web browser for the iPhone and iPod Touch" |
| 01:13 | <AryehGregor> | Who can fail to be impressed by the stunning feature list? I mean, just look at the first (thus presumably most important) one: "iCab Mobile comes with several default search engines you can choose from. You can easily add new search engines. Search suggestions from google and yahoo can be activated. It’s also possible to search within a web page." |
| 01:13 | <AryehGregor> | . . . actually, maybe that is a non-ubiquitous feature for mobile browsers. :( |
| 01:14 | <AryehGregor> | "- Kiosk Mode – a restricted variation of the fullscreen mode for Kiosk environments." For iPhone kiosks? |
| 01:19 | <MikeSmith> | iCab was a lot more impressive before he switched from developing and maintaining his own browser engine |
| 01:20 | <MikeSmith> | I always thought it was pretty amazing that he managed to do that all |
| 01:21 | <MikeSmith> | anybody know what the deal is with Chrome's "integrated PDF viewer"? |
| 01:21 | <MikeSmith> | http://chrome.blogspot.com/2010/11/pdf-goodness-in-chrome.html |
| 01:21 | <MikeSmith> | I mean, what library they are using |
| 01:24 | <robman> | hey MikeSmith - how's it going? |
| 01:24 | <MikeSmith> | oh, and it's also using Pepper? |
| 01:24 | <MikeSmith> | robman: going good |
| 01:25 | <robman> | sorry i disappeared mid conversation |
| 01:25 | <robman> | been swamped all last week 8/ |
| 01:25 | <karlcow> | Now that all big marketing machines are full throttle toward html5… I fear the future of html. I might finally regret Web Services and Co. for attention attractors |
| 01:25 | <MikeSmith> | yeah, last time you were here I think I was asking you what this "trigger" thing was in your slides |
| 01:26 | <robman> | karlcow HTML5 is now almost in the corporate bingo pool 8) |
| 01:26 | <robman> | MikeSmith hrm...the idea was that a pattern language could be used to define mappings from "sensor data bundle" patterns to "web requests" |
| 01:26 | <robman> | loosely described as a triffer |
| 01:26 | <robman> | s/ff/gg/ |
| 01:27 | <robman> | hey...what was that IOT conference in japan you were mentioning too |
| 01:27 | <MikeSmith> | http://www.iot2010.org/ |
| 01:28 | <robman> | cool..t. |
| 01:28 | <robman> | a |
| 01:29 | <robman> | gah...too soon...can't make it 8( |
| 01:30 | robman | looks to see if it's streaming? |
| 01:31 | <MikeSmith> | there's a guy named Dominique Guinard in Zurich who has been doing some interesting work |
| 01:31 | <MikeSmith> | and who will be at that event |
| 01:31 | <MikeSmith> | http://www.webofthings.com/ is one of the things he's involved with |
| 01:32 | <robman> | ta |
| 01:32 | <robman> | oooh...interesting 8) |
| 01:33 | <robman> | hey MikeSmith - is there any work going on to create a single sensors API (e.g. integrating SSN-XG, DAPI, Video/Audio, etc) ? |
| 01:34 | <MikeSmith> | no idea |
| 01:34 | <MikeSmith> | I don't know what SSN-XG or DAPI are |
| 01:34 | <robman> | 8) |
| 01:35 | <robman> | SSN-XG is the Semantic Sensor Network XG and DAPI is the Device API WG |
| 01:35 | <robman> | nevermind 8) |
| 01:35 | <MikeSmith> | ah |
| 01:35 | robman | crawls back under his rock |
| 01:35 | <MikeSmith> | well, my sense is that SSN-XG is going nowhere |
| 01:35 | <robman> | i have heard it's at a "turning point" |
| 01:35 | <MikeSmith> | heh |
| 01:36 | <MikeSmith> | I guess that's a way to describe it |
| 01:36 | <MikeSmith> | I know Laurent Lefort has put a lot of work into it |
| 01:36 | <MikeSmith> | but it does not seem to have generated a lot of interest so far |
| 01:38 | <robman> | 8/ |
| 01:40 | <MikeSmith> | as far as Device APIs WG and work going on with video/audio APIs in other groups, I can't say that we are making a concerted effort at coordination |
| 01:41 | <MikeSmith> | we so far have one instance of an API that started life in the DAP WG but subsequently moved out completely when greater vendor interest in it developed |
| 01:41 | <MikeSmith> | which is the File API |
| 01:42 | <MikeSmith> | and which is now looking like it's on a path to eventually becoming a part of the core Web platform |
| 01:44 | <robman> | hrm...ok |
| 01:47 | <MikeSmith> | in other news, I see Eric Lawrence is still doing regular updates to Fiddler |
| 01:48 | <MikeSmith> | http://www.fiddler2.com/Fiddler/changelog.asp#v2 |
| 05:21 | <MikeSmith> | http://lists.w3.org/Archives/Public/public-html/2010Nov/0061.html was another useful outcome of the meeting |
| 05:22 | <MikeSmith> | oh wow |
| 05:22 | <MikeSmith> | i just noticed this: |
| 05:22 | <MikeSmith> | http://lists.w3.org/Archives/Public/public-html-comments/2010Nov/0000.html |
| 05:23 | <MikeSmith> | from James Clark |
| 05:24 | <MikeSmith> | on how to get an outline for a "chunk" that you want to be the entire content of a page |
| 05:24 | <MikeSmith> | proposing a "nosection" boolean attribute on <body> as possible solution |
| 07:31 | <MikeSmith> | krijnh: I would like to say I meant that tweet sarcastically |
| 07:31 | <MikeSmith> | btw, thanks for tweeting in something other than English |
| 07:31 | <MikeSmith> | or not in English exclusively at least |
| 07:32 | <MikeSmith> | I wish more people would do that |
| 08:14 | <annevk> | Unicode is presentational? Where does that come from? |
| 08:18 | <othermaciej> | Aryeh's message in that regard seems confused |
| 08:19 | <othermaciej> | characters are almost by definition semantic, not presentational, so I would consider a tag equivalent to a character to be semantic |
| 08:19 | <annevk> | yeah, the presentation is in the glyph |
| 08:20 | <annevk> | of course, with some characters the line is somewhat blurry, but still |
| 08:34 | <jgraham> | So, anyone want to summarise what I missed in the last week in about two sentences? |
| 08:36 | <annevk> | sad stuff on ietf-http-wg |
| 08:36 | <annevk> | "I even tried calling Chase customer service about the mixed content |
| 08:36 | <annevk> | warnings on the login page to their banking web site. They told me to |
| 08:36 | <annevk> | use Firefox because the warning didn't show in that browser (this was |
| 08:36 | <annevk> | a case where Chrome's mixed content detector was more accurate than |
| 08:36 | <annevk> | Firefox's in detecting the vulnerability)." |
| 08:36 | <annevk> | though not entirely surprising, my bank until recently advocated IE6 |
| 08:37 | <annevk> | jgraham, reading lots of email |
| 08:42 | <hsivonen> | are all the HTML WG minutes from the TPAC week collected somewhere in the form of links? |
| 08:45 | <annevk> | Laura Carlson made this I noticed: http://www.w3.org/html/wg/wiki/Minutes |
| 08:45 | <annevk> | (or kept it up to date) |
| 08:46 | <hsivonen> | annevk: cool. thanks |
| 08:47 | <hsivonen> | I'm trying to flush my action queue right away |
| 08:48 | <annevk> | I hope I do not have any actions |
| 08:48 | <annevk> | though I guess I should write up something on registries |
| 08:48 | <annevk> | and I got a somewhat indirect request to blog about <bdi> |
| 08:56 | <annevk> | abarth|melting, thanks for the update on URLs |
| 08:56 | <abarth|melting> | annevk: np |
| 08:57 | <abarth|melting> | annevk: hopefully i'll get some time to concentrate on it and make some actual progress |
| 08:57 | <annevk> | abarth|melting, agreed that sorting out hybi deserves priority :) |
| 08:57 | <abarth|melting> | i think the hybi thing is almost done |
| 08:58 | <abarth|melting> | i'm mostly just waiting on some vendors to tell me its ok to release the results of this experiment we ran |
| 08:58 | othermaciej | wonders which vendors are sitting on the results |
| 08:59 | <othermaciej> | annevk: I guess I should send links to each set of minutes |
| 09:01 | <annevk> | I think people have been sending out links already, but maybe a pointer to the wiki on public-html-announce or some such |
| 09:07 | <othermaciej> | yeah, there were some individual links, but not for each session |
| 09:13 | <othermaciej> | this makes me sad: http://news.ycombinator.com/item?id=1880412 |
| 09:13 | <othermaciej> | especially the discussion that blames browsers for flagging mixed content |
| 09:15 | <abarth|melting> | collin had a lot of trouble with one of his web sites |
| 09:15 | <abarth|melting> | making it work over HTTPS without mixed content |
| 09:15 | <abarth|melting> | a bunch of things you want to integrate with don't work |
| 09:15 | <abarth|melting> | like facebook connect |
| 09:15 | <hsivonen> | othermaciej: from the comments, it seems that mixed content warnings are a real problem from the site POV |
| 09:15 | <abarth|melting> | he ended up contracting a CDN to proxy everything over HTTPS |
| 09:16 | <hsivonen> | othermaciej: so it seems at least semi-appropriate to blame browsers |
| 09:16 | <othermaciej> | hsivonen: I'd prefer just not showing the lock instead of showing a mixed-content warning |
| 09:16 | <othermaciej> | (in fact I think that is what Safari does) |
| 09:16 | <hsivonen> | othermaciej: sure. that's what Firefox 4 does, too. |
| 09:17 | <othermaciej> | if sites want to still show a normal HTTPS lock icon on mixed content pages, then there's really no way to help them |
| 09:17 | <othermaciej> | I suppose using SSL with mixed content mostly protects against Firesheep-style passive attacks - you'd have to actually intercept and MITM the non-SSL content to make it still work |
| 09:18 | <othermaciej> | (unless your non-SSL content shares cookies with your SSL content) |
| 09:18 | <abarth|melting> | there's actually a nice way of solving these problems |
| 09:18 | <abarth|melting> | which is to let the site talk TLS to the browser |
| 09:18 | <abarth|melting> | but use the "http" scheme |
| 09:19 | <abarth|melting> | that gives you protection from passive network attackers |
| 09:19 | <abarth|melting> | and doesn't trigger any of these interoperability problems |
| 09:19 | <abarth|melting> | s/interoperability/integration/ |
| 09:20 | <abarth|melting> | if folks are interested, i can write up a document explaining how to do it |
| 09:20 | <othermaciej> | it seems to me that being less noisy about the mixed content warning (just hide the lock and don't show EV cert UI or whatever) would be easier to implement and would serve about as well |
| 09:20 | <abarth|melting> | IE8 will have noisy mixed content errors for a long time |
| 09:21 | <abarth|melting> | (like show a modal dialog noisy) |
| 09:21 | <othermaciej> | is IE8 compatible with the "talk TLS over HTTP" solution? |
| 09:21 | <abarth|melting> | yes, it just doesn't get the improved security |
| 09:21 | <abarth|melting> | :) |
| 09:21 | <othermaciej> | I see, so it would fall back to normal HTTP? |
| 09:22 | <abarth|melting> | it starts out as normal HTTP |
| 09:22 | <abarth|melting> | imagine we magically teleported a policy bit from the server to the client |
| 09:23 | <abarth|melting> | that says "Whenever you see an http URL pointing to me, talk to me at port XXX over TLS instead please" |
| 09:23 | <abarth|melting> | now, all the URLs are http |
| 09:23 | <abarth|melting> | but right before they hit the wire, we decide to send the HTTP requests over TLS instead |
| 09:24 | <abarth|melting> | this is playing on the fact that URL schemes don't have to be the same as on-the-wire protocols |
| 09:24 | <othermaciej> | I was going to suggest this sounds a lot like STS, but I guess STS would specifically not want to allow mixed content |
| 09:24 | <abarth|melting> | its very much like STS |
| 09:25 | <abarth|melting> | but instead of upgrading your https to "bullet proof", you're upgrading your http transport to use tls |
| 09:25 | <othermaciej> | it's just sad that AdSense is one of the integration problems |
| 09:25 | <othermaciej> | but surely there will be others for some time |
| 09:25 | <abarth|melting> | the github guy mentions the charts API |
| 09:26 | <abarth|melting> | presumably Google could clean up its act |
| 09:26 | <abarth|melting> | Facebook connect is a big one |
| 09:26 | <abarth|melting> | i think collin found that he could get that to work if he haxored the script they gave him |
| 09:27 | <othermaciej> | many at Google are pushing for going all-TLS |
| 09:27 | <abarth|melting> | which is basically what the github guy said about the charts API |
| 09:27 | <abarth|melting> | i'm in favor of all-TLS |
| 09:27 | <othermaciej> | I know you are :-) |
| 09:28 | <othermaciej> | I'm just saying it's odd that Google does things like Google Search over TLS while holding out on the API side |
| 09:28 | <abarth|melting> | i'm sure you've heard stories from Microsoft about how no one tells the Office business unit what to do |
| 09:28 | <abarth|melting> | you might imagine that adsense has a similar position :) |
| 09:33 | zcorpan | wonders what correction gsnedders refers to in http://www.w3.org/mid/4CD555C7.8010005⊙oc |
| 09:34 | <othermaciej> | I have heard surprisingly few stories about Microsoft's internal politics |
| 09:35 | <abarth|melting> | brb |
| 09:39 | <abarth> | X-Bodge-Transport-Security |
| 09:39 | <abarth> | i think that turns on that feature in chrome |
| 09:41 | <annevk> | yay more X-prefixed headers |
| 09:41 | <abarth> | annevk: it's meant to be named rediculously |
| 09:42 | <abarth> | i don't think anyone knows it exists |
| 09:43 | <othermaciej> | you just revealed the secret! |
| 09:43 | <abarth> | OMG! |
| 09:43 | <abarth> | actually, it looks like this doesn't quite do what we want |
| 09:44 | <abarth> | http://src.chromium.org/viewvc/chrome/trunk/src/net/base/transport_security_state.h?view=markup |
| 09:44 | <abarth> | MODE_STRICT is for HSTS |
| 09:44 | <abarth> | X-Bodge-Transport-Security gets you into MODE_SPDY_ONLY, which is for SPDY |
| 09:44 | <abarth> | we want MODE_OPPORTUNISTIC |
| 09:45 | <abarth> | but i'm not sure if there's syntax to activate that |
| 09:47 | <abarth> | oh, i see |
| 09:47 | <abarth> | agl is too clever for his own good |
| 09:48 | <abarth> | / TODO(agl): implement opportunistic HTTPS upgrade. |
| 09:48 | <abarth> | :) |
| 09:48 | <abarth> | now that i've gotten this far, I should just implement it |
| 10:09 | <wirepair> | is there something wrong with the html5 sandbox / iframe srcdoc validator on html5.validator.nu? |
| 10:09 | <wirepair> | getting Attribute srcdoc not allowed on element iframe at this point. when taking an example from the html5 spec directly |
| 10:09 | <hsivonen> | wirepair: most likely there is a problem yes |
| 10:09 | <annevk> | might not be implemented yet |
| 10:09 | <wirepair> | ok cool, was scratching my head there :) |
| 10:12 | <wirepair> | anyone know off the top of their heads which browsers have started implementing (in a testable form) srcdoc? |
| 10:16 | <annevk> | I remember some work in progress in WebKit |
| 10:16 | <abarth> | i don't think it's landed yet |
| 11:28 | <annevk> | nice change proposal hsivonen |
| 11:32 | <hsivonen> | annevk: thanks |
| 12:00 | <wirepair> | hu just found a Really Weird ie9 vulnerability |
| 12:00 | <wirepair> | can read remote cssRules *only* when it's framed |
| 12:00 | <wirepair> | w.t.f |
| 12:03 | <zcorpan> | remote as in cross-origin? |
| 12:04 | <wirepair> | yes |
| 12:10 | <david_carlisle> | abarth: on using http: URI scheme rather than https:, seems like you are in good company TBL moans about https appearing in the URI in http://www.w3.org/DesignIssues/Architecture.html#Extension |
| 12:27 | <wirepair> | heh, not only is it only when it's framed, it has to be dynamically framed |
| 12:27 | <wirepair> | wow that's a messed up flow :/ |
| 12:27 | <wirepair> | (dynamically in the sense that the iframe src tag needs to be assigned in javascript, not just in the element src="...") |
| 12:37 | <wirepair> | buh i don't get it, sometimse cssRules are valid, sometimes they're null |
| 12:37 | <wirepair> | even though i see them in their developer tools debugger |
| 12:37 | wirepair | gives up |
| 12:37 | <wirepair> | s/their/there |
| 12:54 | <jgraham> | hsivonen: yt? Did you implement the counters in the AAA as in the spec or only incrementing the inner loop counter whn actually inserting an element? |
| 13:02 | <jgraham> | abarth: ^ |
| 13:08 | <gsnedders> | zcorpan: The final case in jgraham's email which he claimed could never happen |
| 13:09 | jgraham | wonders what he claimed could never happen |
| 13:10 | <gsnedders> | I guess zcorpan was referring to my email to public-html-testsuite |
| 15:27 | <zcorpan> | hsivonen: webkit looks at htmlness? |
| 15:28 | <annevk> | and if there's no ownerDocument you default to XML? |
| 15:28 | <zcorpan> | yeah |
| 15:29 | <annevk> | can you even get in that situation? |
| 15:30 | <zcorpan> | doctypes can have null as ownerDocument :) |
| 15:43 | <hsivonen> | zcorpan: yeah, webkit looks at the HTMLness of the owner doc |
| 15:44 | <zcorpan> | then i agree with your comment that we should align with webkit and add an argument to override |
| 15:45 | <zcorpan> | does anyone support DOMParser(s, 'text/html') ? |
| 15:46 | <annevk> | I think Mozilla does |
| 15:46 | <zcorpan> | uh, that's not the right syntax |
| 15:47 | <hsivonen> | annevk: I'm pretty sure we don't yet |
| 15:47 | <gavin> | our domparser doesn't support text/html yet |
| 15:47 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/694 |
| 15:47 | <gavin> | bug 102699 |
| 15:48 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/695 even |
| 15:48 | <zcorpan> | i guess it's bad that opera doesn't throw |
| 15:48 | <zcorpan> | doc is null in webkit |
| 15:53 | <hsivonen> | spot-checking other Change Proposals, it sees that people rarely list the Negative Effects of their proposals |
| 15:54 | <gavin> | heh |
| 15:56 | <jgraham> | hsivonen: I don't remember anyone else ever filling out the "risks" section of their proposal |
| 15:56 | <jgraham> | It is quite worthless IMHO |
| 16:00 | <zcorpan> | only Hixie putting jokes in the Risks section |
| 16:01 | <hsivonen> | aargh. Wikia sniffs Firefox and Opera and expects their script execution to be HTML5-incompliant |
| 16:08 | <zcorpan> | yay |
| 16:53 | <TabAtkins_> | <!DOCTYPE html><title>SVG in HTML</title><svg><circle cx=10 cy=10 r=10 /></svg> |
| 16:55 | <jgraham> | TabAtkins_: Very nice. But why? |
| 17:03 | <TabAtkins_> | jgraham: Argh, sorry. Unintentional. Copypasta from Hixie's message a day or so ago. |
| 17:09 | <jpike> | yay |
| 17:10 | <annevk> | who, interwebs was down |
| 17:10 | <cdbb> | i've got a question for you pros |
| 17:10 | <annevk> | I mean, whoa |
| 17:11 | <cdbb> | i have a slider, with 10 slides. they are marked up with figure/figcaption. how do i fix my outline (it appears there is a blank list with 10 bullets)? |
| 17:12 | <TabAtkins_> | cdbb: Figures scope their contents from the heading algorithm, so yeah, anything in them won't be shown. |
| 17:12 | <annevk> | <section> seems more appropriate for a slide in a slideshow |
| 17:12 | <TabAtkins_> | If you want them to show in the ouline, instead mark them up with <section>/<h1>. |
| 17:12 | <annevk> | depending on the contents |
| 17:13 | <TabAtkins_> | annevk: Right. If the slides are important enough to be in the outline, they're too important to be <figure>s. |
| 17:13 | <TabAtkins_> | (If they're not, then <figure> is perfect.) |
| 17:13 | <cdbb> | ah |
| 17:13 | <cdbb> | ok, they are a large image with a graphical caption combined into a single image |
| 17:13 | annevk | usually goes for <p> |
| 17:13 | <cdbb> | so i thought figure/figcaption would make the most sense |
| 17:14 | <cdbb> | but section probably makes more sense semantically |
| 17:14 | <cdbb> | thx guys |
| 17:14 | <TabAtkins_> | Np. |
| 17:14 | <annevk> | if it's just a couple of lines i'd go for <h1> and <p><img> for each slide personally |
| 17:14 | <annevk> | <h1> for the title of the presentation |
| 17:15 | <TabAtkins_> | I gave my last talk with <section> slides. The presentation was particularly fun, because they were placed all over the place in the document and used transitions on top/left to animate slide changes. |
| 17:52 | <annevk> | hahaha http://active.tutsplus.com/articles/roundups/html5-and-flash-17-industry-experts-have-their-say/ |
| 17:57 | <annevk> | http://twitter.com/html4nana |
| 17:59 | <othermaciej> | if progress is measured by novelty accounts, HTML5 is doing pretty well |
| 19:26 | <mikekelly> | ok I'm looking at an overview of html5 that says the link element 'lost @type because of the prevelence of css' - is that true? |
| 19:28 | <annevk> | not entirely |
| 19:28 | <annevk> | pointer? |
| 19:29 | <mikekelly> | it's a devzone refcard thing |
| 19:30 | <mikekelly> | annevk: default for link is css ? |
| 19:30 | <annevk> | no |
| 19:30 | <annevk> | for <link> type is mostly meaningless |
| 19:30 | <annevk> | but it's still there |
| 19:31 | <annevk> | see http://simon.html5.org/html5-elements |
| 19:31 | <mikekelly> | sure, where is that article coming from |
| 19:31 | <mikekelly> | and why is it slightly true? :P |
| 19:32 | <annevk> | it isn't really |
| 19:32 | <annevk> | and I've no idea where random articles without URLs on the internet are from |
| 19:32 | <annevk> | or where they get their ideas from |
| 19:32 | <mikekelly> | ok all it says is: "The <link> element has lost its type attribute due to the prevalence of CSS." |
| 19:34 | <mikekelly> | so that's wrong, basically |
| 19:36 | <beowulf> | mikekelly: where on the internet does it say this? |
| 19:39 | <mikekelly> | beowulf: why do you ask? |
| 19:41 | <mikekelly> | @type is a crazily pointless attribute |
| 19:49 | <annevk> | mikekelly, basically, because URL or it didn't happen |
| 19:52 | <AryehGregor> | othermaciej, annevk: Well, I don't want to get into an argument about . . . er . . . semantics. But there's a big difference between semantics in the HTML sense, and typography. Something like a line break or non-breaking space is not semantic by any definition I see -- it specifies appearance and can be used for any semantics where you desire that appearance. Of course Unicode doesn't dictate presentation on a pixel level as CSS sometimes |
| 19:52 | <AryehGregor> | does, either. |
| 19:53 | <annevk> | Unicode does not dictate presentation at all |
| 19:53 | <annevk> | it just defines meaning of code points |
| 19:53 | <AryehGregor> | Some of which are presentational. |
| 19:53 | <annevk> | fonts dictate presentation |
| 19:53 | <annevk> | well sure, Unicode has its shade of gray, just like any other technology |
| 19:55 | <AryehGregor> | Unicode says U+41 is a Latin capital letter A. That suggests how it should be presented, not how it should be used. Unicode doesn't have an opinion on whether you use A as an English letter, a Spanish letter, ASCII art for an Asteroids-style spaceship, or total gibberish. It's semantics-neutral. |
| 19:55 | <AryehGregor> | But it had better look something like Latin capital letter A usually looks, in any font. |
| 19:56 | <AryehGregor> | Otherwise you should be using private-use code points or something. |
| 20:19 | <JonathanNeal> | Is this Section 508 happy all by itself or would it also require a label? <input type="text" name="month" maxlength="2" title="Date of Birth Month" /> |
| 20:20 | <AryehGregor> | You don't need to label all inputs, as far as I know. |
| 20:20 | <AryehGregor> | However, you probably do want to use a label here instead of a title. |
| 20:20 | <AryehGregor> | Titles aren't suited for containing essential info like what you're supposed to put in an input. |
| 20:20 | <JonathanNeal> | I think I'm good, because <input>, <textarea>, and <select> elements have label elements associated with them in the markup or are given a descriptive title attribute. |
| 20:21 | <JonathanNeal> | Hmm. |
| 20:22 | <JonathanNeal> | I'm imaginging this in the context of three inputs representing a date of birth. |
| 20:23 | <Philip`> | How about an <input type="date">? |
| 20:26 | <JonathanNeal> | is there type="year", type="month", and type="day" ? |
| 20:26 | <annevk> | AryehGregor, it being a capital A in itself has meaning |
| 20:26 | <annevk> | AryehGregor, saying it has no semantics is the same as people saying <h1> has no semantics because it is completely unspecific |
| 20:27 | <JonathanNeal> | And I imagine the "Date of Birth" text in ... Date of Birth <input type="text" name="month" maxlength="2" title="Date of Birth Month"/> <input type="text" name="day" maxlength="2" title="Date of Birth Day"/> <input type="text" name="year" maxlength="4" title="Date of Birth Year"/> ... wouldn't be appropriate as a label or as a legend in a fieldset. |
| 20:28 | <Philip`> | (That series of inputs would be extremely confusing to people who aren't used to the US date format, and also confusing to people who try to write "Jan" etc as the month) |
| 20:29 | <JonathanNeal> | Okay. So, 1. a label for each would be appropriate (and not a title)? And 2. the "Date of Birth" text would work best as... ? a heading? |
| 20:38 | <TabAtkins_> | JonathanNeal: Yes to both. |
| 20:40 | <FGRibreau> | Hi |
| 20:40 | <FGRibreau> | is there anyone here ? |
| 20:41 | <TabAtkins_> | Nope. |
| 20:41 | <annevk> | FGRibreau, better to state your actual question ;) |
| 20:42 | <FGRibreau> | I don't know if it's the best place to discuss this. I've read the spec about HTML5 postMessage and I've go a question |
| 20:42 | <FGRibreau> | (and a bug in my case) |
| 20:42 | <FGRibreau> | My parent iframe is on "http://www.webiste.com"; and contain another iframe on "https://www.web2.com"; |
| 20:43 | <FGRibreau> | web2 try a postMessage to "http://www.website.com"; and it doesn't work. |
| 20:43 | <TabAtkins_> | Right. |
| 20:43 | <FGRibreau> | Is it normal ? |
| 20:43 | <annevk> | so you did parent.postMessage() from web2? |
| 20:43 | <FGRibreau> | yep |
| 20:44 | <FGRibreau> | It work when web2 is without https |
| 20:44 | <annevk> | ooh |
| 20:44 | <FGRibreau> | I've tested on safari & chrome |
| 20:44 | <annevk> | https to http? |
| 20:44 | <FGRibreau> | (yes) |
| 20:45 | <FGRibreau> | I thought postMessage was able to deliver a message from https to http & http to https & http to http |
| 20:45 | <annevk> | maybe browsers block that |
| 20:45 | <FGRibreau> | humm |
| 20:45 | <annevk> | in theory yes |
| 20:45 | <FGRibreau> | ok |
| 20:45 | <annevk> | but certain things get disabled when you cross the "secure" boundary and not all of that is defined |
| 20:45 | <FGRibreau> | That was what I was thinking about -__-' |
| 20:46 | <FGRibreau> | thanks for your help |
| 20:46 | <FGRibreau> | I'll try another way to get it work |
| 20:46 | <FGRibreau> | ++ |
| 20:46 | <annevk> | I'd suggest trying Firefox/Opera/IE as well |
| 20:46 | <annevk> | oh well |
| 20:49 | <AryehGregor> | annevk, so let's put it this way: Unicode, like CSS, is not media-independent. It's designed entirely around typography. The concept of a capital A only really makes sense in visual contexts -- in other media (like speech) you have to do some kind of lossy conversion to preserve the semantics. Likewise, <br> and <b> and so on are not media-independent, but <h1> is. |
| 21:05 | <TabAtkins_> | MikeSmith: Um, HTMLWG minutes are being *crazy*. Right now it looks like the minutes are a plain-text dump of a photoshop file. |
| 21:05 | <TabAtkins_> | MikeSmith: http://www.w3.org/2010/11/04-html-wg-minutes.html |
| 21:06 | <annevk> | wfm |
| 21:07 | <TabAtkins_> | wtf |
| 21:07 | <TabAtkins_> | Okay, it works now... |
| 21:19 | <TabAtkins_> | Urgh, now I'm back to the minutes not working at all, which I was having earlier this morning. |
| 21:25 | <TabAtkins_> | annevk: When you pointed me to the bug about an anonymous relation and said something about CSS, are you thinking something in the direction of "background-image: anon-url(http://www.example.com/example.jpg);";? |
| 21:45 | <jamesr_> | what are the recent canvas bugs on w3.org bug tracker talking about? http://www.w3.org/Bugs/Public/show_bug.cgi?id=11242 for example |
| 21:52 | <JonathanNeal> | TabAtkins_, is there an example online somewhere of how to properly use headings, legends, labels, and titles in html5 forms? |
| 21:52 | <JonathanNeal> | And if so, would you share it with me? :D |
| 21:53 | <Philip`> | jamesr_: Accessibility |
| 21:55 | <jamesr_> | so it's about the fallback content inside a canvas? |
| 21:56 | <Philip`> | jamesr_: Yes |
| 21:57 | <jamesr_> | why's canvas different from any other html element? |
| 21:59 | <Philip`> | I don't know |
| 22:00 | <AryehGregor> | jamesr_, because people write entire applications in canvas. |
| 22:00 | <AryehGregor> | Which aren't exposed to AT at all, and can't really be, because canvas is just a bitmap manipulated by JavaScript. |
| 22:01 | <AryehGregor> | I mean, can't really be without special APIs. |
| 22:01 | <AryehGregor> | Whereas other HTML elements tend to contain convenient things like text, which can be exposed to ATs easily. Or otherwise they have semantics that can be exposed to AT. |
| 22:02 | <jamesr_> | right, but people also put entire applications in <embed> |
| 22:02 | <AryehGregor> | Yeah, but the processing model for that is entirely nonstandard, so it's up to the particular embedded thingie to define its own accessibility model. |
| 22:03 | <jamesr_> | so the idea with this series of bugs is that the application author would provide DOM that parallels the application being rendered in the <canvas>? |
| 22:03 | <AryehGregor> | I don't know, I haven't looked at them. |
| 22:06 | <Philip`> | The HTML5 spec already requires that you provide a non-canvas-based equivalent to what you're using canvas for |
| 22:09 | <zcorpan> | maybe i should write a non-canvas-based tower defence game to see what would be suitable as fallback |
| 22:11 | <zcorpan> | i guess it's a bit hard to play tower defence if you're blind |
| 22:40 | <MikeSmith> | TabAtkins_: still have problems with the minutes? |
| 23:10 | <TabAtkins_> | heycam: Yo, in the WebIDL spec there's a reference to the toNumber(), etc. functions, which are claimed to be defined in Chapter 9. Chapter 9 doesn't exist. Thoughts? |
| 23:11 | <heycam> | oh :) |
| 23:11 | heycam | checks |
| 23:11 | <heycam> | ecmascript apocrypha |
| 23:11 | <heycam> | http://sideshowbarker.github.com/es5-spec/#x9.3 |
| 23:12 | <TabAtkins_> | Danke. You may want to update that reference in the webidl spec, because it's pretty confusing. |
| 23:13 | <heycam> | but that's chapter 9 |
| 23:13 | <TabAtkins_> | MikeSmith: Yes. |
| 23:13 | <heycam> | or is the issue that it should say "section 9"? |
| 23:14 | <TabAtkins_> | heycam: The issue is that both me and the dude who asked me the question missed that all the section references in this chapter refer to es5, not to the webidl spec. |
| 23:15 | <heycam> | TabAtkins_, ah gotcha |
| 23:15 | <TabAtkins_> | Because we just searched for "toNumber", and didn't think it was important to read the first paragraph of the chapter that the first hit happened to be in. ^_^ |
| 23:18 | <heycam> | TabAtkins_, btw there is a sentence at the top of that section saying "Section numbers in this section refer to sections of ECMAScript Language Specification, 5th Edition" |
| 23:19 | <heycam> | TabAtkins_, because things were looking messy when i included the target spec for each of those references |
| 23:20 | <MikeSmith> | heycam: Jason Orendorff also has an HTML ES5 spec - |
| 23:20 | <MikeSmith> | http://people.mozilla.org/~jorendorff/es5.html#sec-9.3 |
| 23:20 | <heycam> | TabAtkins_, maybe i could use a different section notation, like ?9.3 |
| 23:20 | <heycam> | MikeSmith, yeah i saw that -- i really like the formatting on that one |
| 23:20 | <MikeSmith> | yeah |
| 23:21 | <heycam> | yours is just the first that came up in my history when i typed "es5" :) |
| 23:21 | <TabAtkins_> | MikeSmith: Yeah, minutes are still causing me problems. I dunno why. |
| 23:22 | <MikeSmith> | TabAtkins_: so what's the problem exactly? it just shows up garbled? |
| 23:22 | <TabAtkins_> | Right now it's just giving me a "this webpage is unavailable" message. |
| 23:22 | <heycam> | TabAtkins_, filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=11268 |
| 23:23 | <TabAtkins_> | For a short period of time, a few hours ago it was barfing out the contents of a photoshop-created image. |
| 23:23 | <MikeSmith> | TabAtkins_: might be a problem with one of the mirrors… lemme check |
| 23:23 | <TabAtkins_> | Woops, now it's back to doing that again. |
| 23:24 | <TabAtkins_> | And now they work again. |
| 23:24 | <TabAtkins_> | I dunno if it's something crazy with me, or what. |
| 23:25 | <TabAtkins_> | Back to "not available". I suspect I pulled from cache a second ago when they worked. |
| 23:27 | <AryehGregor> | So, visiting this URL causes X to crash and my window manager to reset: http://upload.wikimedia.org/wikipedia/commons/thumb/8/8f/Whole_world_-_land_and_oceans_12000.jpg/800px-Whole_world_-_land_and_oceans_12000.jpg |
| 23:29 | <AryehGregor> | No, wait, this is the one that crashed X: http://upload.wikimedia.org/wikipedia/commons/thumb/8/8f/Whole_world_-_land_and_oceans_12000.jpg/800px-Whole_world_-_land_and_oceans_12000.jpg |
| 23:29 | <MikeSmith> | TabAtkins_: mirrors are all OK and I don't see anything weird in the headers for that page or anywhere else that might be causing problems |
| 23:29 | <AryehGregor> | Presumably because it's 19.11 MB. |
| 23:30 | <AryehGregor> | Wait, I mean: http://upload.wikimedia.org/wikipedia/commons/8/8f/Whole_world_-_land_and_oceans_12000.jpg |
| 23:30 | <AryehGregor> | There we go. |
| 23:30 | <AryehGregor> | Don't click it. |
| 23:30 | <TabAtkins_> | MikeSmith: Shrug, then. It seems to load fine initially, then if I refresh it dies, then sometimes if I refresh it starts barfing out an image as text. |
;"){kind=link}
