| 00:00 | <Philip`> | It's also the sound of people rediscovering bugs they'd forgotten about and deciding they shouldn't be allowed to die quietly |
| 00:03 | Philip` | wonders what CCN is |
| 00:03 | <Philip`> | (in relation to http://lists.w3.org/Archives/Public/public-html-bugzilla/2010Mar/0764.html) |
| 00:04 | <Philip`> | Oh, I guess it's http://www.ccnx.org/ |
| 00:05 | <Dashiva> | It refers to http://lists.w3.org/Archives/Public/www-tag/2009Sep/0055.html at least |
| 00:05 | <Philip`> | Hmm, I think I saw Van Jacobson talking about that but I can't remember anything about it :-/ |
| 00:05 | <Dashiva> | "content-centric networking" sounds like a buzzword without much real substance |
| 00:06 | Philip` | should have paid more attention, if it's going to replace IP |
| 00:07 | <Philip`> | Since he's one of the main developers of TCP, I think he knows a bit about the substance of networking |
| 00:09 | <Dashiva> | Lots of failed projects were done by people with excellent technical skills |
| 00:10 | <Dashiva> | These use cases seem to focus on money and control, with a little attribution on the side |
| 00:11 | <Philip`> | I believe it's "content-centric" because you care more about the content than about where it comes from - e.g. instead of sending packets to an IP address, you send requests to a content identifier and it can be satisfied by any router that has cached that content, or else it'll be forwarded towards a more authoritative source for that data |
| 00:12 | <Dashiva> | "We expect that most organizations will have a way to intake GPL code." :) |
| 00:12 | <Dashiva> | They've made a custom binary encoding of XML... |
| 00:14 | <Philip`> | (and you route based on longest-prefix matches of the content identifier, so it can efficiently handle hierarchical naming schemes but isn't strictly constrainted to a hierarchy, so it still gives freedom to separate content from location) |
| 00:16 | <Dashiva> | I wonder how they plan to translate from user desires to interest messages |
| 00:18 | <Philip`> | Users click desirable links on web pages which link to the content name, I'd guess |
| 00:18 | <Philip`> | or programs generate them automatically |
| 00:18 | <Philip`> | just like they do with URLs today |
| 00:19 | <Philip`> | only without today's underlying layers that collapse the whole rich URL structure into a 4-byte address of a host that contains that content |
| 00:20 | <Dashiva> | It seems like they're pulling the search engines into the transport layer |
| 00:21 | <Philip`> | Why does it seem like that? |
| 00:21 | Philip` | sees it as more like routing on URLs (plus caches and security) |
| 00:22 | <Dashiva> | Well, if the search engines remain a necessary part of the system, what is gained from losing the location-dependency? |
| 00:22 | <Dashiva> | Search results already span different locations |
| 00:22 | <Philip`> | (going by the CoNEXT paper at least) |
| 00:22 | <Dashiva> | P2P networks handle distributed downloads too |
| 00:24 | <Dashiva> | I'm watching a video presentation, he says they're NOT Google. That they do the inverse. Google is "content to names", they are "names to content" |
| 00:24 | <Dashiva> | Seems 'content' is rather overloaded as a term here |
| 00:25 | <Philip`> | If you use some extra service like a search engine to map identifiers onto locations, then higher layers don't have to worry about locations and you've got an emulation of location-independence and that should work okay |
| 00:25 | <Philip`> | but it'd be awfully inefficient since it's layered on top of an internet that's fundamentally designed around locations |
| 00:26 | <Philip`> | and you'd have to do a lot more communication than if the underlying layers were designed to support that properly |
| 00:27 | <Dashiva> | Doing a near-broadcast interest message, and then all the possible responders have to manage to alert each other so only one (ideally) responds |
| 00:27 | <Philip`> | P2P works for some things, but it doesn't help distribute the front page of a news site or videos on YouTube |
| 00:27 | <Dashiva> | Well, if you remove the user uploading, you get a CDN |
| 00:27 | <Philip`> | I suppose FreeNode supports distributing web pages but it has the problem of being rubbish |
| 00:27 | <Philip`> | Uh |
| 00:27 | <Philip`> | FreeNet |
| 00:27 | <Philip`> | or whatever it is |
| 00:28 | <Dashiva> | How would this system handle spam |
| 00:28 | <Dashiva> | You could reply to any interest message with Cialis :) |
| 00:30 | <Philip`> | By cryptographically signing content messages so that they're tied to the content name |
| 00:30 | <Dashiva> | But the name is included in the request, so you'd know the name |
| 00:31 | <Philip`> | You map the namespace hierarchy onto a tree of public/private keys |
| 00:31 | <Philip`> | so that only people with the key for a particular name can sign content purporting to be from that name |
| 00:32 | <Dashiva> | How would the keys be distributed and authenticated? |
| 00:35 | <Dashiva> | I suppose you could use a DNS-like system |
| 00:36 | <Philip`> | You trust somebody (perhaps a centralised authority like with DNS) who asserts the validity of other keys, recursively |
| 00:36 | <Philip`> | and keys are just content so they're distributed like any other content (under some kind of naming convention) |
| 00:38 | <Dashiva> | "I tell the server to store _my_ pictures." "I tell the camera the pictures it takes are _mine_" |
| 00:39 | <Dashiva> | Seems like you'd need to tell the server about the camera (or the inverse), or you'd be pinging every machine everywhere for pictures |
| 00:52 | <Dashiva> | "Most communication isn't point to point anymore" |
| 00:55 | <Philip`> | I think one of my doubts was that I expect things are going to become more point-to-point, not less |
| 00:56 | <Philip`> | because everyone is a fan of personalisation and customisation |
| 00:56 | <Philip`> | When you visit a web site you're never going to get precisely the same content as anybody else, because you'll have logged in and it'll be tailored for you |
| 00:56 | <Philip`> | and YouTube will do on-the-fly encoding of targeted adverts into the videos you watch |
| 00:57 | <Dashiva> | This video presentation seems to focus on personal data, all the examples are gadget-related |
| 00:57 | <Philip`> | and the idea of caching content for multiple users will be obsolete |
| 00:57 | <Philip`> | and the idea of requesting the same piece of data twice will also be obsolete |
| 00:58 | <Philip`> | and people will care about services, not about content |
| 00:59 | <Dashiva> | Although perhaps non-personal content (audio, video media) might have big enough in volume to still matter |
| 01:04 | Philip` | sleeps |
| 01:05 | <Dashiva> | They sign the content, yes |
| 01:06 | <Dashiva> | They'll need to include some kind of DRM, though... |
| 01:17 | <Dashiva> | One huge problem seems to be that people are expected to manage (sub)namespaces |
| 01:17 | <Dashiva> | I don't think most people will be able to make sensible hierarchies |
| 01:26 | <jwm> | lots of chatting tonight |
| 01:26 | <jwm> | hehe |
| 01:42 | <Dashiva> | "Security: each packet will be named; the naming will be registered and secure (in the same way that IP addresses are now). " |
| 01:43 | <Dashiva> | If IP addresses are registered and secure, and each packet contains an IP address... |
| 05:47 | <redder> | hi guys & girls |
| 05:49 | <MikeSmith> | redder: hey |
| 05:49 | <redder> | how are you today ? |
| 05:50 | <MikeSmith> | so, regarding http://www.w3.org/Bugs/Public/show_bug.cgi?id=7546#c10, "HTML 5 being superseded by CCN or a similar structural change", anybody know what "CCN" is? |
| 05:51 | <MikeSmith> | redder: I am doing swimingly |
| 05:51 | <redder> | lol |
| 05:51 | <redder> | cool |
| 05:52 | <redder> | i am looking for someone who can design a site for me :) |
| 05:52 | <MikeSmith> | content-centric networking |
| 05:53 | <MikeSmith> | redder: I think this would probably not be the best place to look |
| 05:53 | <MikeSmith> | there are not so many designers here |
| 05:53 | <redder> | any suggestions ? |
| 05:54 | <redder> | i do not know web building |
| 05:54 | <MikeSmith> | redder: how much would you be able to pay somebody to design the site? |
| 06:14 | <MikeSmith> | it seems that CCNx is the more common abbreviation for content-centric networking these days |
| 06:15 | <redder> | MikeSmith |
| 06:16 | <MikeSmith> | yeah |
| 06:16 | <redder> | depends on what they ask |
| 06:16 | <redder> | lol |
| 06:16 | <redder> | i am fair guy |
| 06:16 | <redder> | and what they can offer |
| 06:16 | <redder> | just NO Flash |
| 06:16 | <redder> | lol |
| 06:25 | <othermaciej> | what is CCN? |
| 06:25 | <othermaciej> | as in, what does "content centric networking" actually mean |
| 06:27 | <MikeSmith> | othermaciej: yeah, though not clear how it'd ever "supersede" HTML5 |
| 06:27 | <othermaciej> | whether it would or not, I am now curious what it means |
| 06:27 | <othermaciej> | I seriously have no idea |
| 06:27 | <othermaciej> | it is not a term I have heard before |
| 06:28 | zcorpan | guesses it's the supposedly next term after DHTML, AJAX, HTML5 |
| 06:30 | <othermaciej> | from googling, it sounds considerably lower level than HTML5 |
| 06:30 | <othermaciej> | sounds like it is more a replacement for TCP/IP |
| 06:31 | <MikeSmith> | yeah |
| 06:31 | <othermaciej> | (though I am still not sure what it is exactly) |
| 06:33 | <MikeSmith> | Hixie: I would like to bring http://www.w3.org/Bugs/Public/show_bug.cgi?id=9178 to your attention |
| 06:33 | <MikeSmith> | request for Noah/TAG : add definition of "conforming document" |
| 06:34 | <MikeSmith> | othermaciej: http://www.w3.org/Bugs/Public/show_bug.cgi?id=7034 is one I raised but I'm not quite sure where we're at with it |
| 06:35 | <MikeSmith> | it really seems like a meta-issue, not a request for a specific change to the spec |
| 06:35 | <othermaciej> | MikeSmith: where I'd like to be at with it is one or more bugs that propose specific changes to conformance requirements |
| 06:35 | <othermaciej> | MikeSmith: I don't like having this open-ended bug where everyone commenting has a different idea of what it means |
| 06:36 | <MikeSmith> | yeah |
| 06:36 | <othermaciej> | "remove all document conformance requirememts |
| 06:36 | <othermaciej> | " would be a very clear bug |
| 06:36 | <othermaciej> | but if this bug ever represented that request, then Sam has apparently repurposed it for something different |
| 06:37 | <othermaciej> | Larry also made yet another different request |
| 06:38 | <othermaciej> | and there's all the old history |
| 06:38 | <MikeSmith> | what that bug started out as was 'change "conformance checker" to "ideology checker" or "loyalty checker"' |
| 06:38 | <othermaciej> | yeah, which was kind of a joke |
| 06:38 | <othermaciej> | then at least some people supported removing all document conformance requirements as a serious idea |
| 06:39 | <MikeSmith> | If I were to close it myself at this point, I guess somebody else would likely just re-open it |
| 06:39 | <othermaciej> | now Sam seems to be proposing some set of adding requirements, removing requirements, and making requirements optional |
| 06:40 | <othermaciej> | if you feel Sam's request differs from yours, you could work with him to put it in one or more other bugs, or you could go ahead and let him repurpose the bug |
| 06:40 | <othermaciej> | the biggest thing I'd worry about is if multiple people try to repurpose the bug, all for different purposes |
| 06:46 | <MikeSmith> | I would prefer to remove myself as the reporter of that bug, but I guess there's no way to do actually do that |
| 06:47 | <othermaciej> | not without hacking the database |
| 06:48 | <othermaciej> | btw it looks like moving all those bugs to VERIFIED has been successful at getting issues to move to either CLOSED or REOPENED instead of languishing at RESOLVED |
| 06:48 | <othermaciej> | which makes me happy |
| 06:49 | <othermaciej> | we're up to 325 CLOSED, where yesterday I think we were around 270 |
| 06:50 | <MikeSmith> | I just closed all of the ones I had reported |
| 06:51 | <MikeSmith> | except for two |
| 06:51 | <othermaciej> | I did all of mine except one |
| 06:52 | <othermaciej> | we still have a backlog of 1278 bugs in RESOLVED or VERIFIED |
| 06:53 | <othermaciej> | of which 34 are marked with TrackerIssue or TrackerRequest |
| 06:53 | <othermaciej> | so there's clearly a big backlog of bugs that haven't gotten a reply from the originator |
| 06:53 | <othermaciej> | thanks for doing your part to reduce that number |
| 07:01 | <MikeSmith> | I actually didn't do all mine.. the two I have not closed yet are 7034 (the one we were just discussing) and 9178 (add definition of "conforming document") |
| 07:08 | <JonathanNeal> | http://img710.imageshack.us/img710/2292/smurfperception.png |
| 07:12 | <JonathanNeal> | opps, meant that for the ot |
| 07:18 | <othermaciej> | 9178 is still NEW, so that's not an issue |
| 09:00 | <hsivonen> | is window.location.href supposed to be the empty string while the window is loading? |
| 10:54 | <MikeSmith> | othermaciej: the TAG request for a schema makes we wonder if we should add a new HTML WG bugzilla component for "proposed deliverables" or some such |
| 10:55 | <othermaciej> | MikeSmith: so far our way of adding deliverables has been that someone writes it, and then proposes it as a publication |
| 10:55 | <MikeSmith> | true |
| 10:55 | <othermaciej> | MikeSmith: I'm not sure what we'd do with a proposal for a new deliverable that doesn't come with a commitment to produce it |
| 10:56 | <MikeSmith> | well, in this case, there is already a schema that's been produced |
| 10:58 | <othermaciej> | a commitment to edit it as a WG deliverable then |
| 10:59 | <MikeSmith> | yeah, that makes sense, then |
| 10:59 | <MikeSmith> | that way of describing it, I mean |
| 11:12 | <MikeSmith> | http://lists.w3.org/Archives/Public/public-html-ig-jp/2010Mar/0037.html is a good example of a case where <rb>-less HTML5 Ruby markup makes for easier authoring and code readability than the equivalent way of marking it up in XHTML 1.1 complex-ruby |
| 11:40 | <zcorpan> | MikeSmith: the </rt> tags can be omitted too |
| 11:40 | <MikeSmith> | http://news.cnet.com/8301-30685_3-20000423-264.html |
| 11:40 | <MikeSmith> | "Web guru Tim Bray takes Google Android job" |
| 11:42 | <MikeSmith> | zcorpan: in HTML5 ruby? |
| 11:42 | <zcorpan> | MikeSmith: yes |
| 11:43 | <MikeSmith> | I don't understand.. what would be the semantics of that? |
| 11:44 | <zcorpan> | the same; the rt element is closed at </ruby> in that example |
| 11:52 | <zcorpan> | hsivonen: no (re location.href) |
| 11:52 | <othermaciej> | I was surprised to learn that drawImage() with a "copy" composite operation is no faster than with the default "src-over" operation |
| 11:52 | <othermaciej> | (at least in Safari) |
| 11:54 | <othermaciej> | (drawImage is drawing a canvas that happens to contain all 100% opaque pixels in this case, but I would not expect the implementation to specifically check for that case) |
| 11:57 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Ciframe%3E%3C%2Fiframe%3E%3Cscript%3E%0Aw(window%5B0%5D.location.href)%0A%3C%2Fscript%3E vs http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Ciframe%3E%3C%2Fiframe%3E%3Cscript%3E%0Aw(window%5B0%5D.location.href)%3B%0A%3C%2Fscript%3E |
| 11:57 | <zcorpan> | ...is interesting in chrome |
| 11:57 | <MikeSmith> | zcorpan: about the ruby thing, I misread what you wrote. had thought you were saying the rt element could be omitted |
| 11:58 | <zcorpan> | also if i replace the ; with a space |
| 11:58 | <zcorpan> | weird bug |
| 11:59 | <Philip`> | othermaciej: Drawing something that might contain alpha? |
| 11:59 | <Philip`> | (I know that at least Firefox has separate compositing paths for images that are known to be solid RGB) |
| 12:00 | <Philip`> | ((or Cairo does, anyway)) |
| 12:00 | <othermaciej> | I was testing in Safari, and I believe we don't ever assume a canvas backing store is all RGB |
| 12:00 | <Philip`> | Ah, so doing drawImage(canvas) rather than drawImage(image)? |
| 12:03 | <Philip`> | About ImageData: If the fastest possible unpremultiplying-image-copy code is still slow enough to be a problem, could browsers do higher-level optimisations like having getImageData do a direct copy of the image data and only perform the unpremultiplication if the user accesses .data? |
| 12:03 | <Philip`> | (and then putImageData could do a direct copy with no premultiplying if it knows the user didn't fiddle with .data at all) |
| 12:04 | <Philip`> | (so it can be used for blitting between regions on a canvas or between multiple canvases, doing nothing more computationally complex than memcpy) |
| 12:04 | <othermaciej> | I think that would slow down doing poking at the contents of ImageData |
| 12:04 | <othermaciej> | by at least the branch to do the check for lazy un-premult |
| 12:05 | <othermaciej> | even then, it would be overall slower for cases where you actually do modify the ImageData |
| 12:05 | <hsivonen> | zcorpan: I guess window.location in Gecko is broken, then |
| 12:05 | <Philip`> | I presume anyone who cares about performance would do "var d = imagedata.data; d[i]..." rather than doing the .data lookup every time |
| 12:05 | <Philip`> | even if it's doing nothing more complex than a normal property lookup |
| 12:05 | <zcorpan> | hsivonen: i get 'about:blank' in gecko |
| 12:06 | <othermaciej> | Philip`: so you're assuming the lazy un-pre-multiply would be on accessing the data property? |
| 12:07 | <zcorpan> | hsivonen: or the document's url if it's known when the script runs |
| 12:07 | <othermaciej> | Philip`: it would still mean you have to do the original copy and un-premultiply as a second pass |
| 12:07 | <Philip`> | othermaciej: Yes, so it's fast if you don't look at the data at all, but if you do then it'll have a one-time conversion cost and then pixel access will be as fast as usual |
| 12:07 | <hsivonen> | zcorpan: oh. I get "" for window.location.href and "about:blank" for window.document.documentURI |
| 12:08 | <othermaciej> | Philip`: that one-time conversion cost is going to be a lot for a large image - comparable to the cost of getImageData now, but you have the original cost of the raw copy on top of that |
| 12:08 | <othermaciej> | Philip`: as it is, if you just want to pass image data around and *not* directly poke at pixels, you are better off passing a HTMLCanvasElement than an Image Data |
| 12:08 | <Philip`> | Is the raw copy significant in comparison to that? |
| 12:09 | <Philip`> | Looks like Firefox actually already does a raw copy from the canvas into the ImageData, and then unpremultiplies as a separate stage within getImageData, rather than doing copy+unpremultiply simultaneously |
| 12:09 | <hsivonen> | boohoo. Live DOM saving is broken |
| 12:10 | <hsivonen> | zcorpan: http://bit.ly/9GBt2V |
| 12:10 | <Philip`> | Looks like Firefox also uses a 64KB lookup table for unpremultiplication, which sounds a bit weird to me |
| 12:10 | <othermaciej> | Philip`: that might partly explain why its getImageData is so slow |
| 12:11 | <Philip`> | (https://bugzilla.mozilla.org/show_bug.cgi?id=519400) |
| 12:11 | Philip` | goes away |
| 12:12 | <zcorpan> | hsivonen: seems it's different for data: urls |
| 12:13 | <zcorpan> | hsivonen: about:blank and data: gives empty string in gecko, it seems |
| 12:13 | <zcorpan> | hsivonen: but src=document gives 'about:blank' |
| 12:19 | <hsivonen> | zcorpan: wow. that's weird. |
| 13:32 | <Philip`> | "Roughly speaking 'sce' should be to 'bdi' what <em> is to <i>." - I imagine people have widely varying views on whether that is a positive or a negative |
| 13:33 | <Dashiva> | We need _more_ stuff messing up text directions? |
| 13:35 | <Dashiva> | (And is there a practical solution to the problem of users inserting magic characters that mess up text direction?) |
| 13:39 | <Philip`> | Do you mean magic characters like in http://www.google.com/search?q=%E2%80%AE ? |
| 13:39 | <erlehmann> | Dashiva, what is wrong with a) filtering them out ? or b) enforce the opposite markers after user input ? |
| 13:40 | <erlehmann> | (i can see the problem with option a) |
| 13:40 | <MikeSmith> | sce as a name is even more opaque than bdi |
| 13:40 | <MikeSmith> | what does sce stand for? |
| 13:40 | <jgraham> | enforce * after user input seems to fail for any * |
| 13:40 | <Dashiva> | The problem with b) is that it isn't direction-independent |
| 13:42 | <erlehmann> | indeed :< |
| 13:43 | <Dashiva> | I suppose it's not impossible to build an index of characters that change text direction, but even then you'd be forbidding characters that would be legal in one |
| 13:43 | <Dashiva> | of the directions (or at least harmless) |
| 13:45 | <Dashiva> | http://lists.w3.org/Archives/Public/public-html/2010Mar/0335.html |
| 13:45 | <zcorpan> | Dashiva: there's a POP BIDI OVERRIDE character or similar |
| 13:45 | <Dashiva> | Good thing it's text and not XHTML |
| 13:45 | <Dashiva> | zcorpan: Is that harmless if used in excess? |
| 13:46 | <zcorpan> | i think so |
| 13:47 | <Dashiva> | Although having an equal number of that character to the number of characters in user input seems bad |
| 13:47 | <zcorpan> | oh i guess it might mess things up if used in the middle of words |
| 13:49 | <Dashiva> | As long as you can contain the effects to the user input, the user can only mess up his own content, so that's okay |
| 13:50 | <Dashiva> | Maybe a single POP DIRECTIONAL FORMATTING is enough for any number of modifiers |
| 13:52 | <Philip`> | If the user can mess up his own content, he could use RTL markers in order to write naughty words that bypass profanity filters because they're written backwards |
| 13:54 | <Dashiva> | I don't really worry too much about censorship, but I suppose that's also an issue that needs to be adressed |
| 13:58 | <erlehmann> | Philip`, i routinely bypass wordfilters by using special unicode chars. like đťšťđťš‘đťš’đťšś |
| 13:58 | <erlehmann> | seems that nobody to filter unicode math characters that look like letters |
| 13:59 | <erlehmann> | nobody bothers. my sentence a verb |
| 13:59 | <karlushi> | http://www.tbray.org/ongoing/When/201x/2010/03/15/Joining-Google |
| 14:00 | <karlushi> | "I’m going to have to get savvier about HTML5-based applications, because a lot of smart people think the future’s there, that the “native app” notion will soon seem quaint." -- Tim Bray |
| 14:01 | <MikeSmith> | I hope he's not holding his breath on that one |
| 14:01 | <Dashiva> | When was the last time native apps were on death row? |
| 14:01 | <erlehmann> | but it will be *real soon* (holds breath) |
| 14:01 | <erlehmann> | *cough* |
| 14:01 | <MikeSmith> | every year for the last 5 years |
| 14:01 | <MikeSmith> | soon, very soon |
| 14:02 | <MikeSmith> | the Messiah will return and deliver us all from the evils of Native Apps |
| 14:02 | <Dashiva> | We could have a pool for what happens first: year of linux on the desktop, or death of native apps |
| 14:02 | MikeSmith | snorts green tea out of his nose |
| 14:03 | <MikeSmith> | Dashiva: the year of Linux on the desktop has already arrived |
| 14:03 | <MikeSmith> | it's just that you didn't notice |
| 14:03 | <jgraham> | Well since he is working on non-native apps on a popular linux based consumer device... |
| 14:03 | <MikeSmith> | you need special glasses to see it |
| 14:03 | <erlehmann> | because its a slow process ? |
| 14:03 | <Dashiva> | I think it's possible to predict an increase in non-native apps without the native app dying out |
| 14:04 | <karlushi> | there was a Web site for crystall ball mantra… trying to dig the url |
| 14:04 | <MikeSmith> | I think we are still in the Year of Web Services |
| 14:04 | <MikeSmith> | it's just a very long year |
| 14:06 | <Dashiva> | Well, it's still september, after all |
| 14:06 | <erlehmann> | maybe the WHATWG should get some lobbyists akin to the RIAA so that congress votes to extend the year by 20 years |
| 14:06 | <Dashiva> | What recovery rules would you apply to this? http://lists.w3.org/Archives/Public/public-html/2010Mar/0335.html |
| 14:07 | <erlehmann> | the Mike Smith web services term extension act ;) |
| 14:07 | <Dashiva> | I see two instances of <Rich> without a matching </Rich> |
| 14:08 | <Philip`> | I use the recovery rule that if it's too hard to work out what's a quote and what's new material then I don't bother reading it |
| 14:08 | <MikeSmith> | it's like the "we are currently in a stage of high alert" messages that play continuously in the airport for getting close to 10 years now or whatever |
| 14:08 | <Dashiva> | If the second one is an empty comment, saying "imply </Rich> on encountering an empty line" seems reasonable |
| 14:09 | <karlushi> | aaaah found |
| 14:09 | <karlushi> | http://www.longbets.org/ |
| 14:09 | <erlehmann> | Dashiva, isnt there a parsing specification for that to get a proper DOM out of it ? ;) |
| 14:09 | gsnedders | does backflip |
| 14:10 | <Dashiva> | erlehmann: I'm not aware of any specs that include the <Rich> element |
| 14:10 | karlushi | is surprised there is no bet for 2022 |
| 14:11 | <Philip`> | MikeSmith: Would you prefer them to stop being highly alert, and to allow more unimaginative terrorists onto their planes with bombs? |
| 14:12 | <MikeSmith> | “By 2060 the total population of humans on earth will be less than it is today.” .. "total population" instead of just "population".. that's clearly because it means "net population", and this guy knows about the arrival of the anti-humans who cancel out the population of us normal humans |
| 14:12 | <Dashiva> | Now we have to kill you, MikeSmith |
| 14:12 | <erlehmann> | then … let me see … data state, tag open state, tag name state, data state … wtf am i reading |
| 14:12 | <Dashiva> | I mean, someone else has to kill you |
| 14:12 | <Dashiva> | ... I said nothing |
| 14:13 | <MikeSmith> | Dashiva: you'd have to catch me off-guard first, which will be hard for you guys to do, because I have my special glasses that I can detect what your true appearance |
| 14:13 | <Philip`> | Does "by 2060" mean "on at least point in time between today and 2060"? |
| 14:14 | <karlushi> | there is a missing view of the bets by dates. This month's bets results :) |
| 14:14 | <MikeSmith> | Philip`: no, I would prefer that they be honest and just say, We will forever from now on be in a state of high alert because that works well for us |
| 14:14 | <Philip`> | If so, what about tomorrow? Maybe the random fluctuations in population are stronger than the gradual upward trends, so it'll go down a bit before it goes up again |
| 14:14 | <Dashiva> | MikeSmith: But high alert doesn't work well, by definition. It means more effort than ideal. |
| 14:15 | <Philip`> | By what definition? |
| 14:15 | <Philip`> | If a cake is highly tasty, is that bad by definition because it means more tasty than ideal? |
| 14:16 | <Dashiva> | It's bad for the cake baker |
| 14:16 | <Dashiva> | Since it means more effort to produce the cake |
| 14:16 | <Philip`> | Maybe it means 10% more effort to make it twice as tasty |
| 14:16 | <Philip`> | which would seem to be worthwhile |
| 14:17 | <erlehmann> | The WHATWG is committed to the well being of all participants. Cake and grief counseling will be available at the conclusion of the specification. Thank you for helping us help you help us all. |
| 14:17 | <Dashiva> | Cake... or DEATH? |
| 14:18 | <Dashiva> | http://www.longbets.org/382 |
| 14:18 | <erlehmann> | Curse your sudden and inevitable discovery of the true reason behind all this ! |
| 14:18 | <Dashiva> | Seems like a rather useless bet to make |
| 14:19 | <erlehmann> | „will go to National Rifle Association if Keane wins“ |
| 14:19 | <erlehmann> | so they can shoot the remains ? |
| 14:19 | <erlehmann> | when will the bet be called ? |
| 14:20 | <erlehmann> | ah 2018 |
| 14:21 | <Philip`> | "For the purpose of the bet, Earth should be considered "destroyed" if, at the end of the term of this bet in 2018, zero human beings reside on the surface of the planet." |
| 14:21 | <Dashiva> | http://www.longbets.org/244 |
| 14:21 | <Philip`> | What if we all move into underwater cities? |
| 14:21 | <Dashiva> | Who would the VP be? |
| 14:22 | <MikeSmith> | the Submariner and Aquaman would have to fight for it |
| 14:23 | Philip` | would be sure to sign up as a Big Daddy |
| 14:27 | <erlehmann> | i wanna be mermaid man :> |
| 14:27 | <aednsi> | Do it up |
| 16:47 | <MikeSmithX> | reading http://ajaxian.com/archives/a-better-mobile-web-what-else |
| 16:47 | <MikeSmithX> | proposing -webkit-viewport-fixed |
| 16:49 | <annevk> | would be nice if fixed positioning just worked... |
| 16:50 | <annevk> | position:fixed-i-really-mean-it-this-time ... |
| 16:51 | <MikeSmithX> | position: fixed important!!! |
| 16:51 | <MikeSmithX> | position: fixed dammit! |
| 16:52 | <MikeSmithX> | position: fixed goddammit! |
| 17:03 | <annevk> | we could define a new kind of priority |
| 17:03 | <annevk> | position:fixed !omgiamsonotkiddingthistimearound |
| 17:05 | <Dashiva> | It isn't a priority issue, though, is it? |
| 17:05 | <Dashiva> | It seems like they put a viewport onto the viewport |
| 17:08 | <jwalden> | yo dawg |
| 17:48 | <caio_> | I have a small question. I will can send the <video> contents (from webcam <device>) to a server, through HTML 5 / JavaScript? |
| 17:56 | <AryehGregor> | <device> isn't even fully specced last I checked, let alone implemented. |
| 18:02 | <caio_> | Yes, I Know. There is a way to send the contents of <video> to a server, maybe through Web Sockets? |
| 18:03 | <caio_> | for Web Conferences. |
| 18:07 | <AryehGregor> | If you have some data in JavaScript, you can send it via Web Sockets or AJAX or whatever. |
| 18:07 | <AryehGregor> | But there's no standards-based interface right now to get webcam data, AFAIK. |
| 18:32 | <Philip`> | zcorpan: Ping |
| 18:40 | AryehGregor | wonders why http://tstarling.com/blog/2009/12/response-about-climate-change-action/ is still cached by Google as having his name in the title despite having changed ages ago |
| 18:40 | <AryehGregor> | Second result for my name. Although at least the first is my actual website now. |
| 18:44 | <Philip`> | Maybe because it indexes pages based partly on the text used in links to them, not just the text in the page itself |
| 18:44 | <Philip`> | Oh, but that wouldn't explain the cache |
| 18:45 | <AryehGregor> | Also, I doubt there are any links to that post from other sites at all, let alone ones containing my name. |
| 18:45 | <Philip`> | except, I don't see your name in a cached version of it |
| 18:45 | <Philip`> | so I don't understand the problem at all |
| 18:45 | <AryehGregor> | Search for "Aryeh Gregor". Does that page show up in the top ten results? |
| 18:46 | <Dashiva> | I don't see your name in the title, just in the snippet from the text |
| 18:47 | <AryehGregor> | Fascinating. |
| 18:47 | <Philip`> | It's the 5th page in the list (3rd site in the list) |
| 18:47 | <AryehGregor> | Oh, I get different search results. |
| 18:47 | <AryehGregor> | Due to customization. |
| 18:47 | <AryehGregor> | The customized version has my name in the title, the uncustomized one doesn't. |
| 18:47 | <AryehGregor> | Puzzling. |
| 18:48 | <Philip`> | The first result I get is "Re: Inconsistency in HTML 4.01 regarding NBSP from Aryeh Gregor on ..." |
| 18:48 | <AryehGregor> | Oh, well, I guess I'll just have to wait for that post to fade into obscurity. This is why I use "Simetrical" whenever I'm discussing politics or anything, but Tim Starling didn't know that. |
| 18:48 | <Dashiva> | I didn't know you were Simetrical... |
| 18:48 | <AryehGregor> | Did you see Simetrical elsewhere and not realize I was the same person, or you never heard of Simetrical? |
| 18:48 | <Dashiva> | I suppose that's the point :) |
| 18:49 | <AryehGregor> | The point of using Simetrical these days is mainly to avoid gaming sites and stuff coming up as top hits for my name. |
| 18:49 | <Dashiva> | I think I remember you from #wikia and related places |
| 18:50 | <AryehGregor> | Oh, didn't know you were involved in wiki stuff. |
| 18:50 | <Philip`> | I avoid that kind of problem by having the same name as lots of other people |
| 18:51 | <Philip`> | so nobody can easily tell if a particular person is really me |
| 18:51 | <AryehGregor> | I'm probably the only person in the world named Aryeh Gregor. |
| 18:51 | <Dashiva> | Not much. I co-ran two gaming wikis, and reported the template variables are unavailable to parser functions bug :) |
| 18:51 | <AryehGregor> | Since my first name is extremely Jewish and my last name is extremely non-Jewish. |
| 18:51 | <AryehGregor> | Dashiva, oh, that one. |
| 18:51 | <AryehGregor> | That's an old one. |
| 18:52 | <AryehGregor> | I seem to recall {{#tag:}} is the official workaround, and nobody ever bothered actually fixing it. |
| 18:52 | <Dashiva> | It was properly fixed last year, I believe |
| 18:52 | <AryehGregor> | I avoid templates. :P |
| 18:53 | <Dashiva> | I love templates |
| 18:53 | Philip` | wonders if MediaWiki supports template metaprogramming |
| 18:53 | <AryehGregor> | Sadly, yes. |
| 18:53 | <Philip`> | Is it Turing complete? |
| 18:53 | <Dashiva> | Not quite |
| 18:53 | <Philip`> | Alas |
| 18:53 | <AryehGregor> | I've heard it rumored that it is, but the resource limits are such that it isn't in practice. |
| 18:53 | <Dashiva> | But we did manage to totally overload the job queue on tabularasa |
| 18:53 | <AryehGregor> | Pfft, you don't even have to try to do that. |
| 18:54 | <Dashiva> | Well, we did get some central wikia person posting asking us to do something about it :) |
| 18:54 | <Philip`> | If you care about resource limits, then a desktop PC isn't Turing complete either |
| 18:54 | <AryehGregor> | Wikimedia sysadmins tend to just truncate the job queue if it gets too long. :) |
| 18:54 | <Philip`> | so it's probably best to ignore those limits |
| 18:54 | <AryehGregor> | Philip`, yeah, but there are theoretical limits and practical limits. |
| 18:55 | <Dashiva> | We also ran into the maximum include size |
| 18:55 | <AryehGregor> | Meant to prevent dastardly template metaprogramming from getting out of hand. |
| 18:55 | <Dashiva> | :) |
| 18:56 | <AryehGregor> | Although it doesn't work very well, you can still make pages that take half a minute or more to render on Wikipedia. |
