| 01:31 | <MikeSmithX> | http://www.w3.org/2010/02/rdfa/wiki/RDFa-DOM-API |
| 01:35 | <MikeSmith> | boblet: http://html5.org/tools/web-apps-tracker?from=4846&to=4847 |
| 01:37 | <boblet> | MikeSmith: interesting |
| 01:49 | <boblet> | MikeSmith: was there a particular reason for pinging me on that btw? |
| 01:49 | <MikeSmith> | um, I thought you had asked me about something related to that |
| 01:50 | <MikeSmith> | but I guess it was somebody else |
| 01:51 | <boblet> | not that I recall… :) |
| 05:27 | <MikeSmith> | Hixie: I don't see what http://www.w3.org/html/wg/wiki/ChangeProposals/fixedprefixsimple would give us that we can't already do with the class attribute |
| 07:03 | <wycats> | e |
| 07:20 | <Hixie> | MikeSmith: i don't think we'd have wanted apple to invent a new class name for <canvas> |
| 07:20 | <Hixie> | class names are for authors |
| 07:24 | <MikeSmith> | Hixie: I guess I can't see that Rob's fixedprefixsimple proposal would be much value in practice for anything more author-private use |
| 07:25 | <MikeSmith> | Hixie: is Apple has used this approach for canvas, what would it have looked like? <img -apple-canvas> ? |
| 11:13 | <annevk> | no_mind, we've yet to figure out styling form controls unfortunately |
| 11:14 | <annevk> | though what WebKit is doing for controls in combination with appearance and all seems quite nice, but nobody has written down all the details in a specification yet... |
| 11:14 | <no_mind> | hmm so that spec is yet to be written... |
| 11:15 | <annevk> | pretty much |
| 11:15 | <no_mind> | how one start writing such spec |
| 11:15 | <no_mind> | where ever I give html5 demo, all designers ask this question |
| 11:15 | <annevk> | usually by figuring out the fundamentals |
| 11:16 | <annevk> | e.g. how does form styling currently work across various platforms, what is the effect of setting various properties, etc. |
| 11:16 | <annevk> | then figuring out how appearance affects that, what pseudo-classes are supported, how they all work together |
| 11:17 | <annevk> | and while testing all that writing down implementation criteria for someone that tries to emulate it |
| 11:17 | <annevk> | once we know the fundamentals we can start poking holes and enable new stuff |
| 11:17 | <annevk> | that's pretty much how we're doing most new features that are somewhat based on old APIs |
| 11:18 | <no_mind> | ok |
| 11:18 | <no_mind> | that will mean we have to add new css controls or attributes ? |
| 11:19 | <annevk> | maybe |
| 11:19 | <annevk> | another idea is that appearance:none would disable all styling and you would create the entire control yourself using XBL |
| 11:20 | <annevk> | though we'd need some hooks from XBL into the form API to make that completely feasible |
| 11:20 | <annevk> | e.g. so you can declare a form control matches :out-of-range and things like that |
| 11:21 | <annevk> | anyway, as things stand today CSS is not defined for form controls |
| 11:21 | <annevk> | the CSS WG declared it out of scope and nobody volunteered to fix it |
| 11:24 | <MikeSmith> | XBL2 will save us |
| 11:31 | <Dashiva> | But who will save us from XBL2? |
| 11:32 | <no_mind> | annevk, while leaving other form controls is fine, but styling of calender and time control is very important |
| 11:45 | <asmodai> | wtb: printing of MathML formulae in Mozilla |
| 11:53 | <MikeSmith> | fwiw, I concur with no_mind about calendar-widget styling being important for designers |
| 11:53 | <MikeSmith> | any time I've ever shown a demo of the WF2 input calendar, one of the very first questions is always about whether the style can be changed |
| 11:54 | <no_mind> | MikeSmith, same here. |
| 11:54 | <no_mind> | I think its time to write a specification for this. |
| 11:54 | <MikeSmith> | no_mind: so write one :) |
| 11:55 | <no_mind> | I am confused should this be a separate spec or should we add to an existing HTML5 spec |
| 11:55 | <MikeSmith> | separate spec |
| 11:56 | <MikeSmith> | I'm not myself convinced it's so important to end users |
| 11:56 | <MikeSmith> | I do realize that some designers believe it's importan |
| 11:57 | <MikeSmith> | the alternative for designers who do is just to keep doing whatever they're doing now to provide calendar widgets in their content |
| 12:01 | <no_mind> | not to end users but to designers, it is. Like font-face in css |
| 12:05 | <MikeSmith> | no_mind: well, I think users do actually benefit a lot more from the effects of font-face than they would of calendar-widget styling |
| 12:05 | <klevison> | Whati is best browser to test html5 app? |
| 12:05 | <klevison> | *What |
| 12:06 | <MikeSmith> | no_mind: so getting good font-face support is something that to me at least seems actually pretty important for users |
| 12:07 | <MikeSmith> | klevison: is this a test? |
| 12:07 | <MikeSmith> | we can only pick one? |
| 12:08 | <MikeSmith> | what do I win if I pick the right one? |
| 12:08 | <klevison> | I'm new in this topic |
| 12:08 | <klevison> | I'm developer. |
| 12:09 | <klevison> | But I wanna know which browser I should user to dev. html5 app |
| 12:09 | <MikeSmith> | I guess most people test in multiple browsers |
| 12:09 | <MikeSmith> | including in development/pre-release builds |
| 12:10 | <no_mind> | MikeSmith, how font-support was important while calendar styling is not ? Both are used by designers |
| 12:10 | <klevison> | In your opinion which browser is the best? |
| 12:10 | <MikeSmith> | klevison: lynx |
| 12:10 | <MikeSmith> | no_mind: because it's more important to me, and I consider myself the ideal end user |
| 12:11 | <klevison> | MikeSmith: windows browsers |
| 12:11 | <no_mind> | MikeSmith, okies, from my perspective the calender widget needs to blend with rest of look and feel :) |
| 12:12 | <MikeSmith> | klevison: I've heard that a lot of Windows users tend to use Internet Explorer for some reason |
| 12:13 | <MikeSmith> | if that rumor is true, I'd guess that might be a good target to test your app in |
| 12:13 | <MikeSmith> | no_mind: fair enough |
| 12:40 | <asmodai> | MikeSmith: Given the existence of http://jqueryui.com/themeroller/ I'd say styling is quite important to designers at least. |
| 12:43 | <MikeSmith> | asmodai: I wasn't claiming datepicker styling isn't important to designers |
| 12:44 | <asmodai> | MikeSmith: Ah ok, my bad, must have misunderstood. |
| 12:44 | <Philip`> | zcorpan: Hello |
| 12:45 | <zcorpan> | Philip`: hello |
| 12:45 | <Philip`> | klevison: Browsers all have different levels of support for various features, and different sets of bugs, so none is unequivocally best, and you should test in as many as possible and focus on the ones that your users are most likely to use |
| 12:46 | <Philip`> | zcorpan: I recently remembered that I forgot to do http://philip.html5.org/data/entities-in-attribute-with-no-semicolon-or-alphanumeric.txt |
| 12:47 | <Philip`> | I changed your regexp a bit to exclude > from attribute 'values', because otherwise it seemed to pick up a lot of entities that were in content after an element with attributes |
| 12:47 | <klevison> | Philip`: Is there a ranking of most use browsers? |
| 12:47 | <zcorpan> | Philip`: sweet, thanks |
| 12:48 | <Philip`> | zcorpan: Also I think the file is probably truncated, because it seemed to get stuck at some point (I guess a page hit a worst-complexity case in the pattern) so I killed it after it had stopped for a few minutes |
| 12:49 | <Philip`> | klevison: There are lots of rankings, all slightly different |
| 12:49 | <Philip`> | and it depends a lot on target audience, and country, etc |
| 12:49 | <zcorpan> | Philip`: ok |
| 12:49 | <Philip`> | but http://en.wikipedia.org/wiki/Usage_share_of_web_browsers gives some indications |
| 12:50 | <Philip`> | klevison: The theory is that HTML5 applications ought to work the same in all browsers, so for main development work you should pick whichever one gives you the best development tools (debuggers and network monitors and all that stuff) |
| 12:51 | <klevison> | phil |
| 12:51 | <klevison> | Philip`: ok |
| 12:51 | <klevison> | thanks |
| 13:03 | Philip` | wonder what exciting IE9 features will be revealed today |
| 13:03 | <annevk> | oh, IE9 is announced later today? |
| 13:03 | <Dashiva> | 60% on acid3! |
| 13:03 | annevk | wonders if it will beat Opera Mini 5 in standards compliance |
| 13:04 | <asmodai> | Dashiva: rofl |
| 13:04 | <jcranmer> | SVG and canvas support? |
| 13:05 | <jgraham> | ES5 |
| 13:05 | <asmodai> | Anyone else who does/did MathML? I wonder if an accented variable is done with an <msup> or just as an <mi><mo> pair without <msup>. |
| 13:05 | <jcranmer> | the fastest JS engine? |
| 13:06 | <asmodai> | integrated silverlight? :P |
| 13:06 | <annevk> | we have some (imo) crippled supported for MathML |
| 13:06 | <annevk> | support* |
| 13:06 | <asmodai> | annevk: Opera 5 mini? |
| 13:06 | <asmodai> | err, opera mini 5 |
| 13:06 | <Philip`> | annevk: The existence of IE9 has already been announced, but it'll apparently be discussed in more detail at MIX later today |
| 13:07 | <annevk> | asmodai, we released it earlier today |
| 13:07 | <jgraham> | asmodai: <mover> I think |
| 13:07 | <annevk> | Philip`, always so correct |
| 13:07 | <asmodai> | annevk: Sorry, the mathml support I mean. You meant it's in opera mini 5? |
| 13:08 | <asmodai> | jgraham: Let me try that |
| 13:08 | <annevk> | asmodai, prolly, but also in the normal version of Opera |
| 13:08 | <asmodai> | jgraham: No, not over. |
| 13:08 | <asmodai> | jgraham: That's placing it directly over the variable. |
| 13:08 | <asmodai> | jgraham: what you do with, say, vectors |
| 13:08 | <Philip`> | According to Twitter, "Word has it #IE9 will support #CSS3 and #HTML5" |
| 13:09 | <jgraham> | Philip`: What about Web 3.0? |
| 13:10 | <Philip`> | jgraham: I'd guess they'll only have about 80% support for Web 3.0 |
| 13:10 | <jgraham> | asmodai: Oh you mean like <mi>r`</mi> or so? |
| 13:10 | <Dashiva> | Philip`: As in "someone made this statement up" or, "reliable sources tell us" |
| 13:10 | <asmodai> | jgraham: if you have Mozilla 3.6 with html5 enabled: http://www.in-nomine.org/~asmodai/3d.html |
| 13:10 | <asmodai> | jgraham: See Kijiya's and rendering equation formulae |
| 13:10 | <Rik`> | Philip`: do we know when they will talk about it ? today, tomorrow ? |
| 13:11 | <asmodai> | First is using <msup> and the second is without. |
| 13:11 | <Philip`> | Dashiva: Who knows - the origin is lost in an infinite chain of retweets, and Twitter doesn't encourage accurate attribution and evidence for all statements |
| 13:12 | <jgraham> | asmodai: I think it should be without |
| 13:12 | <asmodai> | jgraham: looks better huh? |
| 13:12 | <Philip`> | Rik`: I think it's the keynote today (and some later sessions), which http://live.visitmix.com/Schedule says is 9am in Las Vegas |
| 13:13 | <asmodai> | oh shoot, need to fix something in that bidirectional |
| 13:13 | <jgraham> | Arguably it should be a combining mark of some sort so it is all one effective character or something |
| 13:13 | <asmodai> | jgraham: I wonder who's authoratative on these kind of issues. |
| 13:13 | <jgraham> | asmodai: Have you looked at the spec? |
| 13:13 | <asmodai> | jgraham: From time to time. But it is not a good way to stay sane :) |
| 13:14 | asmodai | goes to check again |
| 13:14 | zcorpan | sees mathml entities in http://www.in-nomine.org/~asmodai/3d.html |
| 13:15 | <zcorpan> | was it hand-written or generated? |
| 13:15 | <zcorpan> | s/generated/converted from something else/ |
| 13:15 | <jgraham> | asmodai: It looks nice btw |
| 13:16 | <asmodai> | jgraham: Thanks! I am not a designer though, but try to make it clear to read. |
| 13:16 | <asmodai> | Sure as hell goes faster now that we have the firemath addon for firefox. Saves me a lot of time designing. |
| 13:17 | <asmodai> | Designing the formulae that is. |
| 13:19 | <jgraham> | Holy crap, the Stix fonts *still* haven't been released |
| 13:19 | <asmodai> | jgraham: Should get a site update this week. |
| 13:19 | asmodai | stopped poking them whenever that date passed :) |
| 13:19 | <asmodai> | jgraham: http://groups.google.com/group/netscape.public.mozilla.mathml/msg/110fb7f0d1d11c52 |
| 13:19 | <asmodai> | jgraham: Looks like it's raised by default when used as a postfix operator |
| 13:19 | <jgraham> | Well yeah but after hearing they will be ready "real soon now" for the best part of a decade |
| 13:20 | asmodai | grins |
| 13:20 | <asmodai> | Must say that standard Windows 7 with FF 3.6 does work decently for MathML |
| 13:21 | <Philip`> | I like how the Stix web site's front page has a quote from 2002 |
| 13:23 | <jgraham> | I like the timeline best of all |
| 13:24 | <asmodai> | http://www.w3.org/Math/testsuite/mml2-testsuite/Topics/Primes/primes1.xml <-- I am not sure I understand why they switch between ′ and ' or not even using ″ instead of '' |
| 13:24 | <asmodai> | But there they use msup |
| 13:26 | <Philip`> | Isn't a font generally a quite predictable thing to schedule? |
| 13:27 | <Philip`> | You know you need to create n glyphs, and you work out how long that takes based on the zillions of pretty much identical font development projects that have been done before |
| 13:28 | <Philip`> | It's not like they trying to land on the moon or collide hadrons or write a Duke Nukem game |
| 13:39 | <asmodai> | Philip`: Yeah, not sure how they managed to let that go on so long |
| 13:39 | <boblet> | re: Stix, I’d worry about a group making a font when they use a screenshot of an italic font with Windows-level aliasing. does not bode well |
| 13:44 | <Philip`> | boblet: I was going to say they'd probably been using that screenshot since 2002 |
| 13:44 | <Philip`> | but actually it looks like it was added to the site in 2007 |
| 13:44 | <boblet> | Philip`: zing! |
| 13:44 | <boblet> | given creating a typeface is almost 100% details, you’d think they’d be concerned about … well, details like that |
| 13:45 | <Philip`> | Maybe the people who are working on the font have been too busy to update the site, so they got someone else to do it |
| 13:58 | <asmodai> | ah |
| 13:58 | <asmodai> | jgraham: found something: |
| 13:58 | <asmodai> | Not all "mathematical identifiers" are represented by mi elements — for example, subscripted or primed variables should be represented using msub or msup respectively. |
| 13:59 | <asmodai> | Mmm, but that's a primed variable. *sigh* |
| 14:21 | <boblet> | another HTML5 Doctor article: http://html5doctor.com/small-hr-element/ |
| 14:21 | <boblet> | sorry Mike, nothing to top the mother of all O’s from last time |
| 14:22 | <MikeSmith> | boblet: you need a pipe |
| 14:22 | <MikeSmith> | in your photo |
| 14:22 | <boblet> | lol |
| 14:23 | <MikeSmith> | and/or a lab coat |
| 14:23 | <MikeSmith> | all y'all htmldoctors should wear lab coats |
| 14:23 | <MikeSmith> | or whatever the white coats that doctors wear are called |
| 14:23 | <boblet> | yeah, that’s a good idea |
| 14:23 | <MikeSmith> | or scrubs |
| 14:23 | <boblet> | scrubs might be a bit hardcore |
| 14:24 | <boblet> | but doctor/boffin lab coats are a plan |
| 14:25 | <workmad3> | you can get lab coats pretty cheap on ebay :) |
| 14:25 | workmad3 | has an extra large one at home that cost about £5 or £10 |
| 14:25 | <MikeSmith> | or you can just photoshop your heads onto existing images of doctors in lab coats |
| 14:26 | <workmad3> | that would be good :D especially if you did it deliberately badly ;) |
| 14:26 | <workmad3> | or (as a final idea), get a lab coat, and then photoshop the head of a random scrubs castmember onto your own body? |
| 14:27 | <boblet> | hehe. these suggestions are gold! |
| 14:27 | boblet | rolls his eyes when everyone looks the other way |
| 14:27 | <boblet> | :P |
| 14:27 | <workmad3> | I saw that :P |
| 14:32 | <boblet> | curses! ;-) |
| 14:37 | <annevk> | really, someone needs to rewrite CSS |
| 14:37 | <annevk> | CSSOM would work so much better if there was an actual parsing specification that says to create CSSStyleRule objects and all and says to create CSSImportRule objects, that says for @import user agents have to "fetch" style sheets |
| 14:38 | <jgraham> | annevk: Are you volunteering? |
| 14:38 | <annevk> | and that when they are fetched they are to be parsed, and then CSSStyleSheet objects are to be created etc. |
| 14:38 | <annevk> | I have it all in my head, I just don't care about writing it all down because it's so much frigging work |
| 14:38 | <annevk> | just writing down the bits specific to CSSOM is already so much work |
| 14:38 | <annevk> | it annoys me |
| 14:40 | <annevk> | I would be happy to give directions if anyone is interested though |
| 14:49 | <MikeSmith> | annevk: what you are describing is what I had thought the CSSOM spec was |
| 14:49 | <MikeSmith> | before I actually read it |
| 14:50 | <MikeSmith> | I even told other people that's what CSSOM was |
| 14:50 | <MikeSmith> | I guess I should actually read specs before I tell other people what they are specs for |
| 14:51 | <zcorpan> | that'd take away all the fun |
| 14:52 | <Philip`> | MikeSmith: It only becomes a problem if those other people then read the spec, which is unlikely |
| 14:52 | <MikeSmith> | heh |
| 14:52 | <MikeSmith> | touché |
| 14:53 | <annevk> | you thought CSSOM was a CSS parser spec? |
| 14:54 | <gsnedders> | We all know CSSOM is the only spec that matters any more. |
| 14:54 | <annevk> | I have thought of including a section that actually says how CSSOM integrates with the vague notion of CSS parser in the normal CSS specs |
| 14:55 | <MikeSmith> | annevk: for some crazy reason I thought CSSOM probably meant "CSS Object Model" |
| 14:55 | <annevk> | which would say that once you parse an @import rule you create a CSSImportRule object, you fetch the indicated style sheet (unless you decide not to based on the media query provided), etc. |
| 14:55 | <annevk> | MikeSmith, well it does |
| 14:56 | <annevk> | MikeSmith, what I'm complaining about is that the link between object model, syntax, and processing, is missing |
| 14:56 | <annevk> | or links |
| 14:56 | <Philip`> | MikeSmith: You should think of it as analogous to HTML4 + DOM |
| 14:56 | <annevk> | currently we have some notion of syntax which is various CSS specs patched together |
| 14:56 | <Philip`> | which define documents, and a document object model, but don't bother with the chasm in the middle |
| 14:57 | <annevk> | some vague notion of processing (not at all the detail of e.g. HTML5 fetch) |
| 14:57 | <annevk> | and a not so good object model that I'm trying to improve, but I keep running into processing model and syntax to object model mapping issues |
| 14:57 | <MikeSmith> | yeah, the chasm is the part I had kind of assumed was provided instead as a mountain of information the spec |
| 14:59 | <annevk> | I sort of have the feeling that only a few people care about the level of detail |
| 14:59 | <MikeSmith> | this seems similar in lackingness to the HTML4 spec |
| 14:59 | Philip` | 's grammar parser has difficulties towards the end of MikeSmith's statement |
| 14:59 | <annevk> | but they are the people that matter if we want this to work in the same way... |
| 14:59 | <zcorpan> | or XML/DOM |
| 15:00 | <Dashiva> | Argh |
| 15:00 | <asmodai> | boblet: Any St Patrick's Day parades in Osaka? :) |
| 15:00 | <Dashiva> | Now it's <RSS> instead of <Rich>, I need a new parser |
| 15:00 | <MikeSmith> | annevk: if the only few people who are a implementors, then that would seem to make it still pretty important |
| 15:00 | <jgraham> | Dashiva: There are plenty of <RSS> parsers on the net |
| 15:00 | <jgraham> | Try one of them ;) |
| 15:01 | <boblet> | asmodai: nope, although I’m sure any Irish etc pub will be using it for sales |
| 15:01 | <MikeSmith> | but if implementors don't care, then I guess it doesn't matter so much who does |
| 15:01 | <asmodai> | boblet: heh, I'm sure they would |
| 15:01 | asmodai | misses the kansai region |
| 15:01 | <annevk> | MikeSmith, yeah, they're also the people that don't really have much time to give feedback on-the-go |
| 15:01 | <annevk> | meh |
| 15:01 | <boblet> | you used to be here? |
| 15:01 | <Philip`> | At some point there would be few enough people who care about a detail that it takes less effort to get them to converge via bug reports than by writing a spec |
| 15:01 | <boblet> | cool! |
| 15:02 | <asmodai> | boblet: 3 or 4 times by now |
| 15:02 | <annevk> | I'm slowly making progress though; a lot more than two years ago |
| 15:02 | <asmodai> | If all goes according to plan, back in May or so |
| 15:02 | <boblet> | what for? |
| 15:02 | <annevk> | Philip`, yeah, except without a spec it's bad for newcomers |
| 15:02 | <asmodai> | boblet: Making my own okonomiyakis is ok, but still nice to taste the real thing there. |
| 15:03 | <asmodai> | boblet: visiting friends and such. My ex-gf is a Japanese living up in tochigi-ken, near utsunomiya |
| 15:03 | <MikeSmith> | Philip`: I had meant in that sentence to refer to the chasmatic-lackingness-of-necessary-details quality perhaps being similar to the corresponding lackingness in the HTML4 spec |
| 15:03 | <boblet> | へー so you’ve lived here rather than visited huh? |
| 15:04 | <asmodai> | Visited. never lived. |
| 15:04 | <MikeSmith> | Philip`: relative to the contrasting cornucopia of such detail in the HTML5 |
| 15:04 | <boblet> | ping Mike or me when you have dates, as there’s a local HTML5 event in May I could take you to :) |
| 15:04 | <asmodai> | boblet: Although I guess I saw more of Japanese normal life than most people do when they visit Japan. |
| 15:05 | <asmodai> | boblet: I'll keep it in mind :) |
| 15:05 | <MikeSmith> | Philip`: or the my-cup-floweth-over level of detail in the HTML5 spec |
| 15:05 | <boblet> | having a Japanese gf does tend to give you a different perspective than a tourist |
| 15:06 | <asmodai> | boblet: She and her mother were also quite keen on teaching me the reasoning, as far as they knew, behind certain elements of Japanese culture and all that. |
| 15:06 | <boblet> | ping me on Twitter @boblet if I’m not here |
| 15:06 | <MikeSmith> | asmodai: I live in Tokyo but prefer to socialize exclusively with people from Kansai or Kyushu |
| 15:06 | <Philip`> | MikeSmith: I meant the statement before that one |
| 15:06 | <MikeSmith> | or anywhere except Tokyo or Kanagawa-ken |
| 15:06 | <jgraham> | MikeSmith: Do you have a fetish for people from places starting with K? |
| 15:06 | <boblet> | MikeSmith: heh |
| 15:07 | <asmodai> | Of course, I was also brought to the Kumon class of her mother's to help the children with their English pronunciation. And being 1,94m tall is always fun. eeeeeeeeeeeee ooki gaijin da na! |
| 15:07 | <boblet> | that’s pretty huge |
| 15:07 | <asmodai> | MikeSmith: I love kansaiben and the region, maybe it's because I am Dutch or something. |
| 15:07 | <asmodai> | boblet: Good thing my ex is something like 1,75m ;) |
| 15:07 | <MikeSmith> | jgraham: yes, but even more so for people from places that have K in the name but instead the K is pronounced like "sh" or something for some bizarre reason |
| 15:08 | <jgraham> | MikeSmith: That explains so much |
| 15:08 | <MikeSmith> | asmodai, boblet - I'm joking about that, anyway |
| 15:08 | <MikeSmith> | There are actually some nice people who are native Tokyoites |
| 15:08 | <MikeSmith> | or so I have heard, at least |
| 15:09 | <MikeSmith> | I trust that it's true |
| 15:09 | <boblet> | MikeSmith: no, that sounds a pretty accurate description of your ‘liking ppl’ habits to me :P |
| 15:13 | <MikeSmith> | anyway, speaking of sento, it's time for me to go to the sento before it closes |
| 15:14 | <asmodai> | I always find people in Tokyo to be a bit nervous. |
| 15:14 | <wirepair> | hu? it's 12:20... heh |
| 15:14 | <asmodai> | Stressy maybe is more the term. |
| 15:14 | <wirepair> | shouldn't it be closed by now? |
| 15:14 | <asmodai> | wirepair: I guess not in Tokyo ;) |
| 15:14 | wirepair | is in tokyo |
| 15:14 | <wirepair> | ;> |
| 15:15 | <wirepair> | didn't realize they stayed open that late ;> |
| 15:15 | <boblet> | wirepair: Mike’s sento is extra-late |
| 15:15 | <wirepair> | guess so |
| 15:15 | <boblet> | normally last in 11 close and midnight, but that one is like open til 1am or something |
| 15:16 | <boblet> | maybe 1:30 or 2 if Mike’s going now |
| 15:16 | <gsnedders> | MikeSmith just lives weird hours. |
| 15:16 | <wirepair> | hehe |
| 15:16 | <boblet> | sry meant to type ‘last in at 11, close at midnight’ |
| 15:16 | <Philip`> | gsnedders: It's not just his hours that are weird |
| 15:17 | asmodai | chuckles |
| 15:17 | <gsnedders> | Philip`: Have you even laid your eyes upon the weirdness of MikeSmith? |
| 15:17 | <gsnedders> | I went running last time I saw him. |
| 15:17 | <boblet> | lol |
| 15:18 | <boblet> | gsnedders: the ’stache? |
| 15:18 | <gsnedders> | boblet: He didn't have that when I saw him |
| 15:19 | <boblet> | gsnedders: well if you thought he was funny-lookin’ then… |
| 15:19 | <gsnedders> | :D |
| 15:20 | <boblet> | Philip`: how come you don’t have Deja Vu Sans Mono on your Web Font Optimizer? |
| 15:21 | <Philip`> | boblet: I guess I never (thought|bothered) to add it |
| 15:22 | <boblet> | Philip`: come to think of it, how come you aren’t listing them alphabetically *seeing as I like coded that for you and everything*?? |
| 15:22 | <boblet> | [angry] smiley here |
| 15:22 | <Philip`> | Because I haven't updated that site at all since before your feedback |
| 15:22 | <boblet> | or maybe the sideways glance one would be appropriate |
| 15:22 | <boblet> | raspberries to you |
| 15:22 | Philip` | should probably get back to it some time |
| 15:23 | <Philip`> | but I'm hoping if I wait long enough then it'll be obsolete and I won't have to |
| 15:23 | <boblet> | i think i tried installing locally but it needed a hack to some library and i gave up |
| 15:23 | <boblet> | wonder what @mpilgrim is using… |
| 15:23 | <Philip`> | It ought to work with a modern default Perl installation with no fiddling, I believe |
| 15:23 | <boblet> | might give it another go |
| 15:24 | <Philip`> | The only slightly unusual module is something like Digest::SHA which I think is now standard in Perl 5.10 |
| 15:40 | <erlehmann_> | boblet, of course, he is using python ;) |
| 15:41 | <boblet> | Python? aaw man no wonder. that’s soo 3 years ago. He should be using something newer like Go :P |
| 15:42 | <Philip`> | boblet: Did you mean font subsetting? I believe he was using my tool for that |
| 15:42 | <Philip`> | (The offline version) |
| 15:42 | <boblet> | Philip`: yep, and that’s what I guessed tttt |
| 15:42 | <boblet> | hold on, do ppl say tttt for to tell the truth? |
| 15:43 | <boblet> | i think i just made that one up |
| 15:43 | <boblet> | woops |
| 15:45 | <Philip`> | I assumed it was some kind of crazy smiley |
| 15:45 | <boblet> | yep, that sounds like me, crazy smiley man |
| 15:45 | <boblet> | :| |
| 15:50 | <Philip`> | http://ie.microsoft.com/testdrive/ |
| 15:51 | <Philip`> | http://ie.microsoft.com/testdrive/info/ReleaseNotes/Default.html |
| 15:51 | <Philip`> | "XHTML Documents" |
| 15:52 | <Philip`> | "SVG" |
| 15:52 | <Philip`> | "DOM L2 Events" |
| 15:53 | <Philip`> | "The demo of HTML5 <video> at MIX10 was a preview of a future release of the Platform Preview. Stay tuned for more!" |
| 15:55 | <zcorpan> | crap, need vista or 7 |
| 15:58 | <Philip`> | http://blogs.zdnet.com/microsoft/?p=5572 - "Microsoft’s new JavaScript engine is built to take advantage of two cores, with the second core compiling JavaScript down to native machine code to help speed up the browser." |
| 16:00 | <AryehGregor> | So now everyone's using JIT. Awesome. |
| 16:01 | <Philip`> | The preview build license doesn't allow you to publish benchmark results |
| 16:02 | <gsnedders> | Same was true of IE8, no? |
| 16:02 | <annevk> | yeah, that's going to go down well |
| 16:02 | <AryehGregor> | So download it, agree to the license, and get someone else to publish benchmark results who didn't agree to the license. |
| 16:02 | <AryehGregor> | Easy. |
| 16:03 | <Philip`> | That would presumably count as publishing results to the someone else |
| 16:03 | <AryehGregor> | Why? You're not running the benchmark. |
| 16:03 | <jgraham> | Not if they run the benchmarks |
| 16:03 | <AryehGregor> | It doesn't say you can't let anyone use the browser who didn't agree to the license, right? |
| 16:03 | <Philip`> | Feel free to argue that with Microsoft's lawyers :-p |
| 16:03 | <Philip`> | or alternatively disregard the license entirely |
| 16:04 | <jgraham> | My uninformed guess is that they are more interested in discouraging people by threat then in having anything enforcable |
| 16:04 | <jgraham> | *than |
| 16:04 | <AryehGregor> | It's so rare that I'm actually excited about a Microsoft product release. |
| 16:05 | <zcorpan> | is there a vista vm for xp? |
| 16:05 | <erlehmann> | Philip`, what video format are they using ? |
| 16:05 | <AryehGregor> | jgraham, my guess is that the provisions are thrown in just in case they'd be handy, not because they even intend to discourage people. |
| 16:05 | <erlehmann> | zcorpan, drop XP, it's older than my little sister ;) |
| 16:05 | <AryehGregor> | So in the event they do have a lawsuit over something or other, it's one extra claim they can try to get to stick. |
| 16:06 | <AryehGregor> | erlehmann, TeX is older than my big brother, but it's still awesome. |
| 16:06 | Philip` | contributes to Hixie's server load |
| 16:06 | <Philip`> | (55/100) |
| 16:06 | <erlehmann> | :D |
| 16:08 | <Philip`> | "GPU-Powered HTML5" |
| 16:09 | <AryehGregor> | erlehmann, also, there's no <video> support in this version, but they've said there will be in a future version. |
| 16:09 | <AryehGregor> | I assume they'll just use Windows' built-in media stuff, the way Apple uses QuickTime, and they'll support all system codecs. |
| 16:09 | <Dashiva> | Oh ho |
| 16:10 | <AryehGregor> | Which would mean H.264 will be supported by default, but not Theora, unless they add Theora support. |
| 16:10 | <AryehGregor> | AFAICT. |
| 16:10 | <AryehGregor> | But who knows, we'll have to see. |
| 16:10 | <Dashiva> | If IE9 supports XHTML, maybe we can bypass the whole polyglot documents mess |
| 16:10 | <zcorpan> | "The SVG in HTML5 embedding method without using foreign object (i.e. just create an <svg> tag in your HTML file)" |
| 16:12 | <svl_> | "We did see one cool HTML5 feature we weren't expecting: IE9 working using YouTube's implementation of the HTML5 video element." <- means h.264 :( http://arstechnica.com/microsoft/news/2010/03/platform-preview-gives-web-developers-first-taste-of-ie9.ars |
| 16:13 | <AryehGregor> | I think it was obvious that IE9 would support H.264. MS is part of MPEG-LA. The question is whether it would support Theora too. |
| 16:13 | <AryehGregor> | I'd guess not, but you never know what MS will do. |
| 16:14 | <gsnedders> | They already support Vorbis is some things |
| 16:15 | <AryehGregor> | Why is the first picture in the "HTML5" section an Acid3 screenshot, when Acid3 doesn't test any specs that have anything to do with HTML5? |
| 16:15 | <gsnedders> | It does, it has the same author |
| 16:16 | <AryehGregor> | Also, one thing that Acid3 tests is support for downloadable TTFs, which IE has said they'll never support, so they'll never get 100 anyway. |
| 16:16 | <zcorpan> | the links to specs go to dev.w3.org |
| 16:16 | <boblet> | so does "We love HTML5 so much that we want to do it right" mean “2022 suckaz!”? |
| 16:17 | <zcorpan> | i hope they will send spec feedback; implementing a spec usually involves finding bugs in the spec, but i haven't seen much feedback from ms |
| 16:17 | <AryehGregor> | They sent some useful feedback. |
| 16:17 | <gsnedders> | But that'd give away they were implementing it |
| 16:17 | <gsnedders> | And management wouldn't like that |
| 16:18 | <gsnedders> | (oh yay for MS) |
| 16:18 | <AryehGregor> | The IE9 preview actually edges out Fx3.7a2 on SunSpider. Awesome. |
| 16:19 | <AryehGregor> | We are finally arriving at an era where the most recent version of every browser is non-crippled. |
| 16:19 | <boblet> | AryehGregor: and somewhat embarassing for Moz :) |
| 16:19 | <workmad3> | AryehGregor: now we just need to enter the era where all users have the most recent version of a browser |
| 16:20 | <AryehGregor> | boblet, they're already doing a major rework of their JS engine, Jager-whatever it is. |
| 16:20 | <Philip`> | JaegerMonkey |
| 16:20 | <zcorpan> | AryehGregor: unless you get bitten by http://hsivonen.iki.fi/doctype/ie8-mode.png (which will get even more complicated in ie9) |
| 16:20 | <jgraham> | Jägermonkey |
| 16:20 | <Philip`> | The preview build says it has IE5, IE7, IE8 and IE9 rendering modes |
| 16:21 | <boblet> | ya gotta admit, there would have been some happy engineers at Redmond when they finally got a better SunSpider score than a modern browser |
| 16:21 | <zcorpan> | wonder if ie10 will replace the ie9 mode or add another |
| 16:21 | <workmad3> | http://www.taranfx.com/firefox-hardware-acceleration |
| 16:21 | <AryehGregor> | The discrepancy might be because Firefox is falling back to an interpreter in some places. If the blog posts about Jägermonkey are right, it will beat everyone when that's done. |
| 16:22 | <AryehGregor> | Of course, one possible result here is that IE will beat the pants off everyone else by IE10 because Microsoft is pouring billions into it, and then we'll see an IE6-like stagnation. But that seems a lot less likely now, since Google and Apple have a heck of a lot more money than Netscape ever did. |
| 16:22 | <workmad3> | I expect JS performance is going to be a leapfrog race for a year or two until all browsers have superb JS performance :) |
| 16:22 | <AryehGregor> | So, I'm extremely optimistic. |
| 16:22 | <AryehGregor> | Yay! |
| 16:22 | <workmad3> | hell, it's already started |
| 16:23 | <gsnedders> | So their JS engine is called "Chakra" |
| 16:24 | <workmad3> | so it seems :) |
| 16:24 | <AryehGregor> | Yeah, just chart performance over the last couple of years. Two years ago, no one was even using JITs, right? |
| 16:24 | <workmad3> | AryehGregor: don't think so... I think that started with chrome 1 or 2 or something |
| 16:24 | <AryehGregor> | Chrome was released less than two years ago, no? |
| 16:24 | <workmad3> | good point :) |
| 16:24 | <jgraham> | AryehGregor: (without having any particular clue what I'm talking about) I would be pretty surprised if there was another factor of 10 improvement in js peformance |
| 16:25 | <AryehGregor> | September 2008. |
| 16:25 | <workmad3> | for some reason I thought it was back in my uni days, but it wasn't :) |
| 16:26 | <AryehGregor> | jgraham, I imagine that JS in the fastest browsers right now is less than ten times slower than C, so I'd agree. |
| 16:26 | <workmad3> | who knows... maybe comps will start using 'javascript cards' that essentially run JS on chip? :) |
| 16:27 | <workmad3> | but yeah... I doubt a 10-fold improvement over a year or 2 is likely |
| 16:28 | <gsnedders> | I expect significant improvements, still |
| 16:28 | <AryehGregor> | Yes, I agree. |
| 16:28 | <gsnedders> | But I think nothing so radical, and nothing so sudden now |
| 16:28 | <workmad3> | yeah |
| 16:29 | <workmad3> | the only reason such a large improvement was really possible is because JS was pretty much ignored and there was a lot of interpreted language improvements that could be applied very quickly to the interpreters |
| 16:29 | <workmad3> | such as the JIT compilation :) |
| 16:29 | <boblet> | shepazu: so did you have your Mix10 preso already? :) |
| 16:30 | <workmad3> | it was almost along the lines of jumping straight from a java JVM from '95 to one from '05 in only a year :) |
| 16:31 | <JonathanNeal> | they'll support video that's nice of them.l |
| 16:32 | <AryehGregor> | If they support Theora, there's still a chance of it winning the format war. If not . . . oh well. |
| 16:34 | <jgraham> | I would speculate that the best-case realistic possibility is that they support whatever codecs you have installed |
| 16:34 | <AryehGregor> | Yes, but they also control what codecs are installed by default. |
| 16:34 | <AryehGregor> | Same as Apple. |
| 16:34 | <AryehGregor> | So they could choose to add Theora support by default to the OS, or not. |
| 16:34 | <AryehGregor> | Up to them. |
| 16:35 | <Philip`> | http://blogs.msdn.com/ie/archive/2010/03/16/html5-hardware-accelerated-first-ie9-platform-preview-available-for-developers.aspx |
| 16:36 | <gsnedders> | http://ie.microsoft.com/testdrive/benchmarks/Acid3/Default.html |
| 16:37 | <Philip`> | "we’re looking for developer feedback on our implementation of HTML5’s parsing rules, Selection APIs, XHTML support, and inline SVG" |
| 16:37 | <Philip`> | Parsing rules? |
| 16:38 | <gsnedders> | Hmmm... |
| 16:39 | <AryehGregor> | Philip`, so that question means "I'm going to go write up a bunch of test cases to see if IE9 actually implements HTML5 parsing", right? |
| 16:39 | <gsnedders> | What about the html5lib runner? |
| 16:39 | <jgraham> | gsnedders: You read my mind |
| 16:39 | <gsnedders> | http://gsnedders.html5.org/html5lib-tests/runner.html is the most up to date version |
| 16:39 | <gsnedders> | So if anyone has Vista/Win7 at hand :P |
| 16:42 | <boblet> | oyasumi |
| 16:42 | <AryehGregor> | I have one, but it's like two rooms away, so I'm too lazy to test it presently. |
| 16:46 | <Philip`> | gsnedders: http://philip.html5.org/misc/html5lib-runner-ie9.txt |
| 16:47 | gsnedders | wonders how up to date that copy is |
| 16:48 | jgraham | wonders what's with the </B> elements |
| 16:49 | <Philip`> | It seems to *fewer* of those tests when it's in IE8 mode |
| 16:49 | <Philip`> | but only a few fewer |
| 16:51 | <gsnedders> | Philip`: Try running it again, I've just updated it |
| 16:51 | <gsnedders> | (Sorry) |
| 16:51 | <zcorpan> | jgraham: presumably unexpected end tags still create "/B" elements |
| 16:51 | <zcorpan> | so not a full html5 parser |
| 16:54 | <zcorpan> | Test 1 of 24 in data/tests10.dat |
| 16:54 | <zcorpan> | | <svg svg> |
| 16:54 | <zcorpan> | and mathml |
| 16:54 | <zcorpan> | no wait |
| 16:54 | <zcorpan> | they don't parse mathml |
| 16:55 | <annevk> | they support SVG in text/html per HTML5 rules? |
| 16:55 | <Philip`> | gsnedders: http://philip.html5.org/misc/html5lib-runner-ie9.txt |
| 16:55 | <Philip`> | (Updated) |
| 16:55 | <gsnedders> | Philip`: Thanks |
| 16:55 | <gsnedders> | annevk: yes |
| 16:56 | <Philip`> | http://twitter.com/IE/status/10577927335 - "we're going to support the H.264 video codec, it's the de facto standard for video across the broadcasting industry" |
| 16:56 | <annevk> | 10 years from now we'll finally find out whether the IE-was-the-problem advocates of XHTML were right |
| 16:56 | <annevk> | I can't wait |
| 16:57 | <erlehmann> | Philip`, all is lost |
| 16:57 | <annevk> | Philip`, it's also evil |
| 16:57 | <zcorpan> | Test 17 of 24 in data/tests10.dat - they don't break out of foreign lands :-/ |
| 16:57 | <erlehmann> | maybe the EU should do a default codec ballot screen :P |
| 17:01 | <Philip`> | Going by the Zombie DOM Viewer, <foo><bar/><baz> is parsed with XMLish nesting |
| 17:01 | <AryehGregor> | Can someone tweet @IE asking if they'll support Theora in addition to H.264, or if they're committed to not supporting Theora? |
| 17:01 | <Philip`> | i.e. unknown tags are no longer empty elements |
| 17:01 | <Philip`> | and trailing slashes make them closed |
| 17:01 | <annevk> | o_O |
| 17:01 | <erlehmann> | AryehGregor, as if they would. |
| 17:01 | <AryehGregor> | erlehmann, we could ask. |
| 17:03 | <Philip`> | Unmatched closing tags of known elements ("</a>" etc) turn into empty elements named "/A" etc in the DOM |
| 17:03 | <zcorpan> | Test 21 of 24 in data/tests10.dat - they don't parse xlink:href into a namespaced attribute :-/ |
| 17:03 | <Philip`> | (</div> and </p> turn into <div> and <p>, though) |
| 17:03 | <Philip`> | (and </unknown> is dropped) |
| 17:04 | <erlehmann> | AryehGregor, do it, then. i'm not on twitter. |
| 17:04 | <AryehGregor> | I'm not either. |
| 17:04 | <AryehGregor> | Someone here must be. |
| 17:04 | <JonathanNeal> | http://twitter.com/jon_neal |
| 17:04 | <JonathanNeal> | there you go AryehGregor. |
| 17:05 | <AryehGregor> | I wouldn't have phrased it that way, but okay. |
| 17:05 | <JonathanNeal> | What's wrong with the way I phrased it? I can always tweet again. |
| 17:05 | <annevk> | then get your own twitter account AryehGregor |
| 17:05 | <annevk> | :) |
| 17:05 | <JonathanNeal> | I don't take twitter seriously enough to worry how I look. |
| 17:06 | <AryehGregor> | It sounds more confrontational than I'd have said it. |
| 17:06 | <Philip`> | http://hsivonen.iki.fi/test/svg-and-mathml-in-html.html displays the nice tiger |
| 17:06 | <zcorpan> | if they are changing the parser in radical ways, one wonders why they don't just implement it per spec in one go |
| 17:06 | <AryehGregor> | I'd have said something like "Might IE9 support Theora in addition to H.264, or is that out of the question?" |
| 17:08 | <erlehmann> | AryehGregor, I would assume that professional PR team behind that twitter account will pack their „no theora 4 u“ message into a nice, fluffy package regardless. |
| 17:08 | <AryehGregor> | What a nicely cynical prediction. |
| 17:08 | <AryehGregor> | They might just ignore it, though. |
| 17:17 | <JonathanNeal> | erlehmann is right, which is why I really didn't care :) |
| 17:17 | <JonathanNeal> | They don't care about me, I don't care about them, it's just a call for information |
| 17:19 | <annevk> | http://www.zeldman.com/2010/03/16/ie9-preview/ is arguing for convergence on WebKit |
| 17:20 | <annevk> | I don't see why Zeldman does not realize that one rendering engine would be quite bad as well... |
| 17:22 | <AryehGregor> | It's always nice to have competition. |
| 17:22 | <JonathanNeal> | hey paul_irish! |
| 17:22 | <jgraham> | Fortunately there is no realistic possibility of Microsoft switching to WebKit |
| 17:22 | <erlehmann> | actually, as a non-corporate person, i wonder why. |
| 17:23 | <erlehmann> | shouldn't „instant better browser“ be a management metric ? |
| 17:23 | <AryehGregor> | Not if it means you can never beat some of your key competitors in certain areas, because they get all your improvements. |
| 17:23 | <AryehGregor> | WebKit is largely LGPL. |
| 17:24 | <AryehGregor> | Using code from Chrome, now, that might make sense, but it would look bad for Microsoft. |
| 17:24 | <AryehGregor> | Probably they figure the extra time and money is worth the benefits to MS's image. It's already widely viewed as a company that just steals good ideas from others. |
| 17:24 | <AryehGregor> | Without having to actually take *code* from others. |
| 17:25 | <Philip`> | Microsoft would have its own ideas about how the browser architecture should be developed in order to provide the best experience for their users (e.g. scrap all the cross-platform layers and make it heavily optimised for Direct2D etc), which are likely to conflict with the goals of Apple's WebKit developers |
| 17:25 | <Philip`> | so they'd have to fork it instead of contributing upstream |
| 17:25 | <AryehGregor> | That's true too. They're integrating heavily with Windows-specific stuff. |
| 17:26 | <AryehGregor> | I think it probably makes sense for them to work on their own. |
| 17:26 | <Philip`> | which would mean they lose the benefits of sharing future development effort from other people, and they have to fix all bugs themselves in a codebase which they have no experience with |
| 17:52 | <roc_> | "scrap all cross-platform layers" doesn't really buy you that much |
| 17:53 | <AryehGregor> | Surely it does if you can ask the Windows team to add any new APIs you'd like. |
| 17:54 | <AryehGregor> | (not that I'd know, I don't ever do platform-specific code) |
| 17:54 | <asmodai> | mm |
| 17:55 | <Philip`> | Asking the Windows team to add APIs doesn't seem a great strategy, because (a) they'll probably say no, and (b) you'll have to wait five years for it to be released and adopted before you can start relying on it |
| 17:55 | <asmodai> | http://ie.microsoft.com/testdrive/ |
| 17:56 | <AryehGregor> | Philip`, I'm talking about the IE team asking the Windows team to add APIs. |
| 17:56 | <Philip`> | AryehGregor: So am I |
| 17:56 | <AryehGregor> | And backport to Vista if necessary. |
| 17:56 | <jgraham> | Philip`: Well IE9 needs things relased since IE8 |
| 17:56 | <jgraham> | *released |
| 17:56 | <AryehGregor> | Why do you think either (a) or (b) is the case? Since MS views IE as part of its core platform, you'd think it would tell the Windows and IE teams to cooperate. |
| 17:57 | <erlehmann> | seems they politely ignored the theora question. |
| 17:57 | <AryehGregor> | Also, hasn't MS historically used undocumented APIs in some of their programs? |
| 17:57 | <AryehGregor> | That's the same sort of idea. |
| 17:57 | <jgraham> | s/things/windows versions/ |
| 17:58 | <AryehGregor> | I'm pretty sure the various MS teams coordinate with each other to a considerable degree. Various things Raymond Chen has said tell me that the compiler team does things at request of the Windows team, for example. |
| 17:58 | <AryehGregor> | Or at least in coordination with them. |
| 18:00 | <asmodai> | Wasn't Opera Mini 5 released out of beta? |
| 18:00 | <gsnedders> | Yes |
| 18:00 | <gsnedders> | Today |
| 18:01 | <asmodai> | Mmm, guess the android market place is not updated yet. |
| 18:01 | <gsnedders> | Sometime this morning before I got to the office :P |
| 18:01 | <asmodai> | Heh |
| 18:02 | <asmodai> | If I install the beta from the market place, will it get the release? |
| 18:02 | gsnedders | claims no knowledge of Mini |
| 18:02 | <asmodai> | Heh |
| 18:02 | <gsnedders> | Provided the rendering engine works… |
| 18:02 | <asmodai> | I assume they'll replace the beta with the release version on the market place. Lets just install it |
| 18:07 | <Philip`> | AryehGregor: I don't know how closely teams cooperate, but it seems there's a fair amount of independence |
| 18:07 | <AryehGregor> | That kind of statement is hard to disagree with. |
| 18:07 | <Philip`> | e.g. "The OS team decides what OS features to update in patches. The hardware team decides what hardware to release. The two teams do not necessarily have the same priorities, especially since the OS team views the hardware team as "just another hardware vendor."" (http://blogs.msdn.com/oldnewthing/archive/2008/08/06/8835316.aspx) |
| 18:07 | <AryehGregor> | Hardware probably isn't as important to MS as IE is. |
| 18:07 | <MikeSmith> | hats off to the IE product-dev team |
| 18:08 | <Philip`> | and I'd expect that to some extent the IE team is seem as just another software vendor (though a very important one, but not necessarily more important than all third-party vendors) |
| 18:09 | <MikeSmith> | this is real progress and always good to leave the audience wanting more |
| 18:09 | <annevk> | given that Trident is used by other Microsoft software and Windows itself wouldn't it be more important? |
| 18:09 | <AryehGregor> | MS views IE as a key part of Windows, so I'd say it's probably viewed as a lot more important than most vendors. |
| 18:09 | <Philip`> | The IE-favouritism and undocumented API things got Microsoft into a little bit of legal bother, so I'm not sure they want to do much more of that |
| 18:09 | <AryehGregor> | They can still release documented APIs. Probably no one will sue them too much for that. |
| 18:10 | <roc_> | it's not good to leave the audience wanting <canvas>, though :-) |
| 18:10 | <MikeSmith> | annevk: as far as the webkit-convergence think, I would think that most people would be smart enough to never take advice from a guy with a mustache, or a hat, or worse yet a mustache and a hat.. because that guy is clearly trying to hide something |
| 18:11 | AryehGregor | has both a mustache and a hat. |
| 18:11 | <AryehGregor> | Although I have a beard too, so I don't know if it counts. |
| 18:11 | <MikeSmith> | beard is a sign of wisdom |
| 18:11 | <MikeSmith> | prophets must have beards |
| 18:11 | <MikeSmith> | nobody takes them seriously otherwise |
| 18:12 | <MikeSmith> | but how many prophets can you think of who just had a mustache? |
| 18:12 | <annevk> | IE9 does not have <canvas>? |
| 18:12 | <annevk> | weird |
| 18:12 | <Philip`> | No(t yet) |
| 18:13 | <AryehGregor> | Why is it weird? |
| 18:13 | <jgraham> | What's the point of a fast js engine if you can't play js games? |
| 18:13 | <AryehGregor> | Most of the stuff they're adding seems like it's higher-priority than <canvas>. |
| 18:13 | <AryehGregor> | Performance, <video>, SVG. |
| 18:13 | <Philip`> | They'll apparently update the platform preview every 8 weeks or so |
| 18:13 | <AryehGregor> | Heh. |
| 18:13 | <AryehGregor> | Point. |
| 18:13 | <MikeSmith> | hopefully it's just the first beta and canvas will be added shortly |
| 18:13 | <AryehGregor> | It's not a beta, it's pre-beta. |
| 18:13 | <Philip`> | and maybe they'll add new features then, when they're more confident they'll definitely be able to ship them |
| 18:13 | <AryehGregor> | But that would be awesome, yes. |
| 18:14 | <Philip`> | jgraham: It might be useful for, like, useful things instead |
| 18:14 | <MikeSmith> | AryehGregor: is your beard a serious manly beard, or more like a billy-goat beard? |
| 18:14 | <Philip`> | like email and office applications |
| 18:15 | AryehGregor | wonders why the heck he gets pictures of naked women when Googling "beard" |
| 18:15 | <Philip`> | Personalised search? |
| 18:15 | <AryehGregor> | . . . |
| 18:15 | <othermaciej> | I would have thought <canvas> is higher priority than SVG given author demand |
| 18:15 | <AryehGregor> | My beard is kind of like this, just a half-inch trim overall: http://en.wikipedia.org/wiki/File:JoeyHerrington.png |
| 18:15 | <AryehGregor> | Except darker-colored. |
| 18:16 | <AryehGregor> | SVG is older than <canvas>, I guess, and better known, and part of Acid3 . . . dunno. |
| 18:16 | <AryehGregor> | Personally, I skimmed the canvas section of the spec, because I don't understand all this crazy graphics stuff. |
| 18:17 | <jgraham> | Philip`: Typically email and office applications aren't really limited by raw execution performance afaict |
| 18:17 | <MikeSmith> | that beard is too tame.. beards should be wild and unruly and awe-inducing |
| 18:17 | <jgraham> | (obviously some applications could be but the obvious examples would also want <canvas>) |
| 18:18 | <AryehGregor> | C'mon, benchmarks are the only reason to improve JS performance. Doesn't everyone know that? |
| 18:18 | <Philip`> | jgraham: I assume they occasionally are limited by JS performance, for a short period of time that is long enough to cause visible latency, which is a good enough reason to optimise JS execution |
| 18:19 | <Philip`> | (even if most of the time they're idle or network-limited) |
| 18:20 | <MikeSmith> | here's a dude with substantial beard: |
| 18:20 | <MikeSmith> | http://offthemeatrack.files.wordpress.com/2007/11/mel4.jpg |
| 18:21 | <MikeSmith> | he looks a bit sad and confused, though |
| 18:21 | <MikeSmith> | I guess sometimes prophets gets sad and confused too |
| 18:21 | <MikeSmith> | anyway, we do need canvas |
| 18:22 | <jgraham> | Philip`: There was a microsoft research paper suggesting that they were more likely to be limited by DOM + Events |
| 18:22 | <MikeSmith> | hopefully some soonly upcoming non-pre beta will add that |
| 18:22 | <jgraham> | Although I didn't think the paper was that great |
| 18:23 | <jgraham> | (I mean it was fine, but the conclusions were not that surprising) |
| 18:23 | <AryehGregor> | Someone can try tweeting them to ask if they're planning on <canvas>. |
| 18:23 | <gsnedders> | You'll get no answer |
| 18:23 | <gsnedders> | It's futile |
| 18:23 | <AryehGregor> | Also, if more people harass them on if they'll support Theora in addition to H.264, maybe they'll answer that. |
| 18:23 | <AryehGregor> | gsnedders, they answered a video codec question. |
| 18:23 | <AryehGregor> | You never know. |
| 18:24 | <Philip`> | That question was already answered in published material |
| 18:24 | <Philip`> | I don't think they're going to release any new information via Twitter |
| 18:24 | <AryehGregor> | Hmm, where? |
| 18:24 | <AryehGregor> | I didn't see it. |
| 18:25 | <Philip`> | http://blogs.msdn.com/ie/archive/2010/03/16/html5-hardware-accelerated-first-ie9-platform-preview-available-for-developers.aspx - "we showed HTML5 video support at the MIX conference, as well as how HTML5 video (specifically industry standard HD-encoded, H.264 720p) [...]" |
| 18:25 | <miketaylr> | here, i believe, AryehGregor: http://twitter.com/IE/status/10577927335 |
| 18:25 | <miketaylr> | oops |
| 18:25 | <AryehGregor> | Philip`, point. |
| 18:25 | <AryehGregor> | Okay, then, Twitter is probably pointless. |
| 18:25 | <AryehGregor> | (as usual) |
| 18:27 | <MikeSmith> | or if we want it to become true, we could just do I want things to become true, which is, I just add them to Wikipedia with a citation to some site I control where I emphatically affirm that it's true, quoting "sources that asked to remain unnamed in order to protect their confidentiality" |
| 18:27 | <AryehGregor> | Revert, personal website, does not meet [[WP:RS]]. |
| 18:28 | <Philip`> | Then you wait until a news site cites the wiki article, and then you make the wiki article cite the site |
| 18:28 | <MikeSmith> | it worked for the Judith "Weapons of Mass Destruction" Miller and the New York Times |
| 18:28 | <AryehGregor> | I wish I had a good response to that, but I don't, because it's actually happened. |
| 18:28 | <AryehGregor> | The moral for me is that Wikipedia shouldn't cite journalists on subjects where you could conceivably cite anything more authoritative. |
| 18:28 | <Philip`> | You can verify the consistency of statements, but you can't verify the truth |
| 18:28 | <MikeSmith> | it's trivial to do, despite the editards |
| 18:29 | <Philip`> | and it's easy to set up consistent loops of untruths |
| 18:29 | <MikeSmith> | yes |
| 18:29 | <MikeSmith> | and easy to base a whole career on them |
| 18:29 | <AryehGregor> | Philip`, I'm fairly sure that I can verify the truth of the statement "Manhattan is not currently a desolate wasteland after a surprise nuclear strike by China this morning." |
| 18:30 | <MikeSmith> | witness Charles Pellegrino and Henry Holt and the Last Train to Hiroshima saga |
| 18:30 | <Philip`> | AryehGregor: You can only verify it by determining that it is inconsistent with your observations |
| 18:30 | <MikeSmith> | I hope China does that just to rid us forever of a setting for Sex and the City |
| 18:31 | <erlehmann> | MikeSmith, then it would continue like CSI … SatC: Miami |
| 18:31 | <AryehGregor> | Philip`, however, I can verify the truth of the observation its negation would be inconsistent with, insofar as I'm in Manhattan right now and can directly observe that it's not a desolate wasteland. |
| 18:31 | <AryehGregor> | (at least, not the whole thing; haven't checked downtown lately) |
| 18:32 | <miketaylr> | as long as they keep the fallout from my neighborhood in southern brooklyn, i'm okay with that scenario as well |
| 18:32 | <MikeSmith> | Miami is high on the list too.. when I was in Texas and watched Armageddon when it first came out, everybody cheered with the aliens destroyed Houston |
| 18:32 | <roc_> | personally I would have prioritiized canvas higher than SVG |
| 18:32 | <roc_> | it's a lot easier, and probably more used on the Web |
| 18:33 | <Philip`> | Is it used by any of the top 7000 sites? |
| 18:34 | <MikeSmith> | the waste toxicity of the east river provides a barrier that would protect Brooklyn for a Manhattan attack |
| 18:34 | <miketaylr> | truth. |
| 18:34 | <MikeSmith> | roc_: I would rather they'd done both |
| 18:34 | <roc_> | I kind of expect they will |
| 18:35 | <MikeSmith> | yeah |
| 18:35 | <roc_> | Philip`: I don't know, but I bet SVG isn't |
| 18:35 | <Philip`> | I doubt they'd ask so many questions about the spec if they didn't have some intention of implementing it at some point |
| 18:36 | <Philip`> | They've probably asked more questions about canvas than about all the parts of HTML5 they've already implemented and shipped or announced, combined |
| 18:36 | <Philip`> | (Admittedly that's still only about five questions) |
| 18:36 | <MikeSmith> | maybe their canvas needs more work and they judged that having a partial SVG implementation was OK but that waiting to have a non-partial canvas would be better |
| 18:36 | <othermaciej> | Philip`: I believe Google Maps can optionally use <canvas> |
| 18:37 | <roc_> | Google Maps can use SVG too |
| 18:37 | <gsnedders> | That is true |
| 18:37 | <roc_> | implementing canvas on top of Direct 2D is really quite easy |
| 18:37 | <Philip`> | Depends if you care about all the edge cases |
| 18:37 | <Philip`> | (I'd guess) |
| 18:38 | <roc_> | maybe, but existing browsers don't get all the edge cases right either |
| 18:38 | <hsivonen> | If google maps can use SVG, why doesn't it? |
| 18:38 | <roc_> | it would certainly be easy to get something going for a "tech preview" |
| 18:39 | <roc_> | hsivonen: it does, it uses SVG for the routes in Gecko, last I checked |
| 18:39 | <hsivonen> | roc_: ah. not for the map, though |
| 18:39 | <roc_> | for the actual map tiles, I expect SVG is a giant performance lose |
| 18:39 | <othermaciej> | I think it is capable of using CSS transforms and transitions for the map tiles now |
| 18:41 | <hsivonen> | does the IE9 preview run on XP? |
| 18:41 | <Philip`> | No |
| 18:41 | <hsivonen> | are they doing video using Vista/7-only APIs? |
| 18:41 | <Philip`> | Yes |
| 18:42 | <Philip`> | They're seemingly doing all rendering using Vista/7-only APIs |
| 18:42 | <hsivonen> | that's an interesting decision |
| 18:42 | <Philip`> | (Direct2D) |
| 18:42 | <hsivonen> | I guess I need to purchase another Windows license in order to poke at IE |
| 18:43 | <Dashiva> | I suppose it's time to try out DX 10+ too... |
| 18:43 | <Philip`> | Unsurprisingly, it doesn't appear to work in Wine |
| 18:43 | <miketaylr> | hsivonen: i found out the hard way: http://miketaylr.com/post/dbf12ce8.png |
| 18:44 | <roc_> | that's probably a good decision for them |
| 18:44 | <Philip`> | (http://ie.microsoft.com/testdrive/info/FrequentlyAskedQuestions/Default.html mentions XP) |
| 18:44 | <miketaylr> | ah yes, FAQS. i've heard of those. |
| 18:44 | <no_mind> | why doesnt microsft puts IE to rest |
| 18:45 | <Philip`> | "Will Internet Explorer 9 run on Windows XP? It’s too early to talk about features of the Internet Explorer 9 Beta." |
| 18:45 | <Philip`> | They're not saying it *won't* run |
| 18:45 | <Philip`> | so they could add a non-hardware-accelerated rendering mode, perhaps |
| 18:46 | <Dashiva> | (I think people sticking to IE6 are a bigger problem than XP users on IE8, overall) |
| 18:47 | <hsivonen> | does D2D work on VirtualBox? |
| 18:49 | <Philip`> | IE9 works for me in VMware on Linux |
| 18:49 | <roc_> | I don't know |
| 18:49 | <roc_> | VMWare does support D2D, accelerated even |
| 18:49 | <Philip`> | so it's not getting real hardware access there |
| 18:50 | <roc_> | VMWare does support some level of hardware acceleration |
| 18:50 | Philip` | has no idea what layers perform emulation in this case |
| 18:51 | <Philip`> | My hardware doesn't really support much acceleration, and I'd be shocked if the OS/driver/X11/VMware/etc compatibility was good enough for it to reach all the way through to a VM :-) |
| 18:52 | <hsivonen> | well, I guess VMWare has the record-and-replay bonus, although it's not relly productized on Linux yet |
| 18:53 | hsivonen | just decided against purchasing Windows 7 plus VMWare two weeks ago. |
| 18:53 | Philip` | is using the free VMware Player so it doesn't do anything fancy like that |
| 18:54 | <Philip`> | For some reason it doesn't even like transmitting mouse-up events if the cursor is not in the top-left few hundred pixels of the screen, either |
| 18:54 | <Philip`> | which makes resizing windows tricky |
| 19:08 | <Philip`> | http://msdn.microsoft.com/en-us/ie/ff468705.aspx |
| 19:08 | <Philip`> | "Added HTML5 support |
| 19:08 | <Philip`> | * Improved parsing of generic elements |
| 19:08 | <Philip`> | * Improved parsing of overlapping tags |
| 19:08 | <Philip`> | * script and style block parsing changes |
| 19:08 | <Philip`> | * Text selection APIs" |
| 19:14 | <Philip`> | "Internet Explorer Platform Preview follows the HTML5 draft spec and resolves overlapping tags at parse time." |
| 19:25 | <zcorpan> | "* script and style block parsing changes" - wonder what they changed. looking at html5lib tests they haven't implemented the double escaped dash dash stuff |
| 19:25 | <gsnedders> | No reparsing? |
| 19:25 | <gsnedders> | But if they have the old stuff from the spec… |
| 19:25 | <gsnedders> | Well, I guess they have their compat modes |
| 19:27 | <zcorpan> | i think they still reparse if i interpret the output correctly |
| 19:27 | <Philip`> | zcorpan: It says later on |
| 19:28 | <Philip`> | "Internet Explorer 8 and earlier versions do not pass text, in either a script or style block, to text nodes in the DOM. Internet Explorer Platform Preview persists text in script or style blocks to the DOM as text nodes, as expected. This enables you to display source code on your page using script." |
| 19:28 | <zcorpan> | Philip`: ok. i don't see that in the html5lib test output though |
| 19:29 | <zcorpan> | oh wait |
| 19:29 | <zcorpan> | i was looking at tests without a doctype |
| 19:30 | <zcorpan> | that explains it |
| 19:49 | <Dashiva> | "Bray is the co-inventor of the XML Web standard" |
| 19:49 | <Dashiva> | Does "invent" make sense for something like XML? |
| 19:51 | <erlehmann> | Dashiva, what would you say ? That he discovered the fundamentals of XML ? ;) |
| 19:51 | <gsnedders> | I invented a subset of SGML! |
| 19:51 | <erlehmann> | That's nothing ! I invented a subset of ASCII containing only the letter x ! |
| 19:52 | <erlehmann> | xxxxxxxxxxxxxx … see ? |
| 19:52 | <gsnedders> | Wow. |
| 19:52 | <gsnedders> | Impresssive. |
| 19:53 | <Dashiva> | erlehmann: I would say designer, editor, or even architect |
| 19:53 | <Dashiva> | Inventor sounds wrong to me |
| 19:53 | <zcorpan> | what about Namespaces? |
| 19:55 | <erlehmann> | Dashiva, fine. I hereby proclaim myself the architect of lowercase-x-notation ! |
| 19:56 | <erlehmann> | to collect royalties, x-LA will be set up soon ;) |
| 19:57 | <erlehmann> | i think inventor is fundamentally right, even if it may ring the wrong bells here or there |
| 19:57 | <Dashiva> | Namespaces are evil enough that I can imagine a mad scientist inventing them |
| 19:58 | <erlehmann> | you mean a mad engineer |
| 19:58 | <erlehmann> | unless there's a hypothesis hidden in the spec |
| 19:58 | <erlehmann> | tag soup parsers are the control group ! |
| 20:00 | <Dashiva> | The hypothesis was that people would actually use it |
| 20:01 | <erlehmann> | mad science indeed |
| 20:15 | <annevk> | Hixie, your response on the progress element question does not really answer the question |
| 20:35 | Philip` | sees a Q&A video with Dean Hachamovitch |
| 20:36 | <Philip`> | in which the second question the host asks is about canvas support |
| 20:37 | <Philip`> | and he doesn't say anything specific but notes that all of IE9's graphics are hardware-accelerated, and that there are more things to come than have been announced |
| 22:14 | <zcorpan> | so ie9 won't need the html5 shiv |
| 22:18 | <othermaciej> | nice |
| 22:18 | <othermaciej> | they have correct handling of unknown elements? |
| 22:18 | <zcorpan> | no, <foo/> gets self-closed |
| 22:19 | <jgraham> | http://research.microsoft.com/en-us/um/people/cormac/papers/2009/SoLongAndNoThanks.pdf seems like it might be interesting |
| 22:21 | jgraham | notices the text of the license on that paper forbids "redistribution to lists", wonders what that means and whether posting a URL counts |
| 22:23 | <Philip`> | Summary: "Microsoft recommends leaving users ignorant to simple security vulnerabilities" |
| 22:24 | <Philip`> | although, in the interest of not entirely blatantly misrepresenting it, I suppose I should note that it ends with "this paper is not to be read as an encouragement to end-users to ignore security policies or advice. The opinions expressed are those of the author." |
| 22:25 | <Philip`> | (Also that was my summary, not a direct quote) |
| 22:25 | <jgraham> | Philip`: I haven't read it all yet but it doesn't seem to say that, really |
| 22:26 | <jgraham> | It notes that it is somewhat rational to ignore security advice if the cost of understanding and following that advice is greater than the expected loss from not following that advice |
| 22:26 | <Philip`> | Sure, but worrying about what it really says is not the way to write headlines on tech news sites |
| 22:27 | <jgraham> | Which seems like a major over simplification since you need to look not just at the expected loss but at the distribution of possible losses |
| 22:27 | <Philip`> | and for most people the cost of reading the actual paper is less than the benefit they gain from doing so |
| 22:27 | <jgraham> | e.g. if you can have all your money stolen that is really bad even if there is only a 1 in a million chance it happens to you and so the expected loss is < $1 |
| 22:28 | <Philip`> | Why is that really bad? |
| 22:28 | <Philip`> | Or, (why) is it more bad than two million people each guaranteed to lose < $1? |
| 22:30 | <jgraham> | I guess the total possible effect on your life scales more than linearly with the amount of money lost (beyond a certian point) |
| 22:35 | <Philip`> | jgraham: Only up to a certain point |
| 22:35 | <Philip`> | with that point being when you are bankrupt |
| 22:35 | <Philip`> | and then you don't care any more |
| 22:35 | <jgraham> | (the other problem with this typ of calculation is that the asumption of rational behaviour is totally flawed. In particular people are probably totally incapable of estimating the probability of a given attack, so if they happen to make "rational" decisions it is probably happenstance) |
| 22:37 | <Philip`> | (Incidentally, would you suggest it is rational to play the lottery, because super-linear scaling means a 10^-7 chance of winning £1M outweighs a 10^0 chance of losing £1?) |
| 22:38 | <Dashiva> | Philip`: No, the value of the million is subject to decreasing value since it's so big |
| 22:38 | <Philip`> | The abstract sounds ambiguous about the suggested behaviour |
| 22:38 | <roc> | you can argue that due to various kinds of poverty traps, playing the lottery is rational in some cases |
| 22:39 | <Philip`> | i.e. whether it means user behaviour (driven by ignorance and laziness) happens to coincide with rational behaviour, or whether it's actually suggesting users are making rational evaluations themselves |
| 22:39 | <Dashiva> | I think the paper should also be read as implementor-facing: "These behaviors are bad for users, so don't make your security depend on them" |
| 22:40 | <jgraham> | Dashiva: That was the interesting point of view for me |
| 22:40 | <othermaciej> | is risk-aversion natural? |
| 22:40 | <othermaciej> | er rational? |
| 22:40 | <othermaciej> | most people dislike losing what they have more than they like gaining something |
| 22:40 | <Dashiva> | only if they understand the risk, though |
| 22:40 | <roc> | it depends |
| 22:41 | <roc> | I think the main point of the paper is that we tend to not rationally consider the costs of security measures |
| 22:41 | <othermaciej> | it's a well-known result in behavioral economics: http://en.wikipedia.org/wiki/Endowment_effect |
| 22:41 | <roc> | we == vendors and researchers |
| 22:42 | <roc> | asking everyone in the world to do something that takes only a few seconds is actually a tremendous cost |
| 22:42 | <roc> | that's a humbling thought for browser developers :-) |
| 22:42 | <othermaciej> | otoh I don't think endowment effect applies in the case of a very low probability of a large loss |
| 22:42 | <othermaciej> | this is why I hate conformation dialogs as a security measure |
| 22:43 | <othermaciej> | even if everyone read it and carefully considered the implications, the cost would likely still exceed the benefit |
| 22:43 | <roc> | people are irrationally fearful of certain kinds of losses, like shark attacks |
| 22:43 | <Philip`> | When there was a 2 minute silence in America for some people who had died, I wondered how many lifetime's worth of productivity was being lost because of that |
| 22:43 | <Philip`> | But it was only about 20 years of productivity, so it wasn't even a whole person |
| 22:43 | <othermaciej> | people are scared of low-probability risk of death or serious injury, if the scenario is particularly vivid |
| 22:44 | <othermaciej> | that's another well-known cognitive bias |
| 22:44 | <roc> | so there's the answer |
| 22:44 | <roc> | "if you don't look at this site's certificate, you will be eaten by a shark" |
| 22:44 | <Dashiva> | Philip`: Sort of like that rant about emails sent to large lists |
| 22:44 | <othermaciej> | however it doesn't tend to apply to risks where the consequences seem commonplace |
| 22:44 | <othermaciej> | especially where costs of avoiding the risk would require ongoing active effort |
| 22:45 | <Dashiva> | (For certificates, pretending it's http seems like a great solution to me) |
| 22:47 | <othermaciej> | I think the paper ignores the fact that certain security measures have a very high false positive rate only because they exist |
| 22:47 | <othermaciej> | for example, if there was no error for invalid certs, a lot more phishing sites would use them |
| 22:47 | <jgraham> | I was going to point out the same thing |
| 22:47 | <othermaciej> | they only avoid using them because there is a scary warning |
| 22:48 | <Dashiva> | But as long as the sites weren't flagged as valid certs, it would be no different from now in effeciency |
| 22:48 | <roc> | othermaciej: is that actually true |
| 22:48 | <roc> | ? |
| 22:49 | <othermaciej> | roc: I haven't seen the full counter-factual world of high degree of phishing/pharming but no warning whatsoever on invalid certs |
| 22:49 | <othermaciej> | but it may mean the level of scariness of the warning is nearly irrelevant |
| 22:49 | <roc> | right |
| 22:50 | <othermaciej> | roc: I do know that at least some malware modifies your local DNS settings and system root certs |
| 22:50 | <Philip`> | http://news.netcraft.com/archives/2010/03/05/new_browser_reports_over_half_of_ssl_sites_may_be_unsafe.html |
| 22:50 | <Philip`> | "A new internet browser, Comodo Dragon, reports that more than half of the world's valid SSL certificates are unsafe." |
| 22:50 | <othermaciej> | whence I conclude that it is at least somewhat valuable to attackers to be able to appear 100% like the real site, if they were able to do so |
| 22:50 | <Philip`> | Also: "Although Comodo states that many websites connected to cyber-crimes use domain-validated certificates, Netcraft's phishing site feed shows that only 0.3% of reported phishing sites use HTTPS, including those running on compromised servers with SSL certificates already in place." |
| 22:51 | <othermaciej> | phishing from a non-SSL server is extremely effective |
| 22:52 | <othermaciej> | it's hard to tell how much marginal impact there is on the success rate of phishing from measures that make it harder to look exactly like the real site in all ways |
| 22:52 | <roc> | right |
| 22:53 | <othermaciej> | the people aware enough to benefit from the protection were already at much lower risk of phishing in the first place |
| 22:54 | <roc> | you might be right that phishers would be more likely to use invalid certs if there was no warning, but I haven't seen any evidence that that would actually make a difference to users |
| 22:58 | <Dashiva> | <a href="www.evil.com">www.PayPal.com</a> |
| 22:58 | <Dashiva> | Nice example :) |
| 23:00 | <Philip`> | It'd be a nicer example if they hadn't forgotten the http:// |
| 23:05 | <Dashiva> | You could draw a connection from cost-vs-benefit of security to the otherwise normal practice of insurance |
| 23:05 | <Dashiva> | It's not irrational to pay more than expected average risk if it lets you avoid large worst case risks |
| 23:17 | <othermaciej> | indeed, but it depends on the magnitude and likelihood of the risk |
| 23:20 | <Dashiva> | The paper also ignores non-monetary costs borne by users |
| 23:21 | <Dashiva> | If you get phished, the bank doesn't just give you the money right away, there's probably many hours of calling and talking and writing to recover it. |
| 23:27 | <AryehGregor> | Philip`, for an example of the Windows team doing stuff for the IE team, look at Vista's Low Integrity mode. I've seen a lot of stuff that implies that was intended mainly for IE -- official MS sources tend to mention low-integrity mode and protected IE in the same breath. |
| 23:28 | <AryehGregor> | jgraham, the usual assumption of economists is that rational agents should attempt to maximize expected utility. Of course, utility doesn't map linearly to money, but the actual distributions don't matter, only the expectation (of utility). There are pretty good theoretical reasons for this; if you don't do this, you can be money-pumped (at least sometimes). |
| 23:30 | <AryehGregor> | Also, people often do tend to act rationally, and not just by chance. Your average person hasn't been hit by online fraud, no one they knows has, and they notice that no one else they know seems to be very concerned, so they (correctly) conclude that it's not a big problem for them. |
| 23:32 | <AryehGregor> | Any way you slice it, it's clear that things like raising a big scary warning for cert errors are just stupid. |
| 23:32 | <AryehGregor> | It's security charade. |
| 23:40 | <Dashiva> | Dang, only 55 on Acid3 |
| 23:40 | <Dashiva> | My prediction was 60 |
| 23:43 | <AryehGregor> | It will probably exceed 60 by the final release. |
| 23:44 | <AryehGregor> | Possibly even by the beta. |
| 23:44 | <Dashiva> | My prediction was for MIX, though |
| 23:45 | <AryehGregor> | Oh. Then you fail. Pretty close, though, good job anyway. |
| 23:49 | <zcorpan> | "if you don't look at this site's certificate, you will be eaten by a shark" - LOL |
| 23:51 | <Dashiva> | It's a trick, the shark is sneaking up behind you while you read it |
