| 00:18 | <hsivonen> | I was reading about externality taxing in the CO2 context on the plane |
| 00:19 | <hsivonen> | the economics of it seem applicable to inaccessible Web pages |
| 00:19 | <hsivonen> | except tracking tax on those would be more difficult |
| 00:21 | <jwalden> | "on the plane"? |
| 00:21 | <jwalden> | er |
| 00:21 | <hsivonen> | jwalden: while flying back from Dublin |
| 00:21 | <jwalden> | nm me |
| 00:21 | <jwalden> | misparsing |
| 00:21 | <jwalden> | was binding "the CO2 context on the plane" as one unit |
| 00:22 | hsivonen | notes that Tim Harford writes about "moral posturing" in the absence of externality taxing |
| 00:23 | <hsivonen> | threatening people with lawsuits is a kind of an externality tax |
| 00:23 | <hsivonen> | but an uncertain one that applies only to big companies |
| 00:23 | <jwalden> | a hideously less efficient one, to be sure |
| 00:23 | <jwalden> | courts and lawsuits epitomize deadweight loss |
| 00:23 | <hsivonen> | yeah |
| 00:24 | <hsivonen> | I'm not a fan of lawsuit threatening as part of the accessibility advocate argument set |
| 00:24 | <othermaciej> | I think it would be hard to justify a large tax on inaccessible web pages on a Pigovian basis |
| 00:25 | <othermaciej> | in fact pure economic analysis might find the cost of making web pages accesible exceeds the likely economic benefit |
| 00:25 | <othermaciej> | I'm also not sure lack of accessibility can be modeled as a negative externality |
| 00:25 | <othermaciej> | rather, it is failure to provide a positive good to some people |
| 00:26 | <othermaciej> | pollution is a negative externality |
| 00:26 | <othermaciej> | giving free candy to some people but not others isn't |
| 00:26 | <hsivonen> | othermaciej: I agree that presence should be considered a positive thing, but you can't start subsidising anyone with an inane Web site for making it accessible |
| 00:28 | <othermaciej> | current approaches to accesibility in law involve neither subsidy nor tax |
| 00:28 | <othermaciej> | rather they rely on legal notions of public accomodation, and impose regulations on anything considered such |
| 00:29 | jwalden | is reasonably happy with how that's fallen out so far |
| 00:29 | <othermaciej> | well, it does create unintended consequences |
| 00:29 | <hsivonen> | I thought in the U.S. they relied on punitive damages acting as a probabilistic externality tex |
| 00:29 | <hsivonen> | tax |
| 00:29 | <jwalden> | I'm not sure what actually motivates orgs at the end of the day |
| 00:30 | <othermaciej> | hsivonen: the same could be said of most regulations |
| 00:30 | <jwalden> | government business is at least part of it |
| 00:30 | <hsivonen> | othermaciej: indeed |
| 00:30 | <othermaciej> | however the risk of damages does create unintended consequences |
| 00:30 | <othermaciej> | for instance, since the ADA passed, unemployment among the disabled is up |
| 00:30 | <hsivonen> | yeah, 508-based government sales are the carrot |
| 00:31 | <jwalden> | on the other hand, I seem to recall it was this channel that mentioned <http://www.nytimes.com/2008/01/20/magazine/20wwln-freak-t.html?_r=1&oref=slogin>; once |
| 00:31 | <othermaciej> | government business seems like a better motivator |
| 00:31 | <gavin_> | hsivonen posted that link, I remember |
| 01:27 | <Hixie> | i love the guy who said that having <p> was bad because it wasn't machine checkable and proposed <div role=paragraph> as a solution |
| 01:27 | <othermaciej> | lol |
| 01:27 | <Lachy> | which mailing list was that on? |
| 01:29 | <hsivonen> | Lachy: public-html and wai-xtech |
| 01:29 | <Lachy> | found it |
| 01:31 | <Hixie> | vlad wants 'textStyle' instead of 'font', for consistency with <canvas> APIs instead of CSS |
| 01:31 | <Hixie> | opinions? |
| 01:31 | <Lachy> | I did some major restructuring of selectors api. It's now a lot more organised, but still have to fix up the nsresolver requirements |
| 01:32 | <Philip`> | Hixie: Makes sense to have 'text' in the name, for consistency with every other text-drawing property and method |
| 01:32 | <Philip`> | and also makes it clearer that it's a CSS style thing, not just a font name |
| 01:33 | <othermaciej> | Hixie: would the value of 'textStyle' be a font name? |
| 01:34 | <Hixie> | othermaciej: it would be the css 'font' property |
| 01:34 | <othermaciej> | I think it should be called 'font' |
| 01:34 | <othermaciej> | most graphics APIs describe the way to draw text with a "font", not a "text style" |
| 01:35 | <othermaciej> | whereas they do have notions like "line style" or whatever |
| 01:35 | <othermaciej> | so I do not buy the consistency argument |
| 01:35 | <Hixie> | ok in that case please reply to vlad's e-mail :-) |
| 01:35 | <Philip`> | People seem to use CSS properties like 'font-family' far more than they use 'font', so the analogy between the canvas API and CSS doesn't really work since people aren't familiar with 'font' |
| 01:36 | <othermaciej> | right, I forgot that 'font' is a shorthand |
| 01:36 | <othermaciej> | it would be more appropriate to provide individual properties for the pieces of that shorthand, for canvas |
| 01:36 | <othermaciej> | IMO |
| 01:36 | <Hixie> | ew |
| 01:36 | <Hixie> | we don't do that with fillStyle |
| 01:37 | <Lachy> | yeah, I think it should stay as font, for consistency with CSSOM property |
| 01:37 | <othermaciej> | Hixie: but fillStyle is not a shorthand, it's just overloaded for different types |
| 01:37 | Philip` | imagines very few people use the font CSSOM property, so that doesn't seem a compelling reason |
| 01:39 | <othermaciej> | Hixie: I will have to do some study to give a cogent reply |
| 01:39 | <Lachy> | Philip`, it's fairly common to see people change styles by using element.style.xxx = ...; for a whole variety of styles. |
| 01:39 | <Lachy> | although I can't say how frequently each property is used |
| 01:40 | <Philip`> | othermaciej: So instead of saying "ctx.font = '10pt Arial'" when you want to set the font and aren't sure of the current canvas state, you'd have to say "ctx.fontStyle = ctx.fontVariant = ctx.fontWeight = 'normal'; ctx.fontSize = '10pt'; ctx.fontFamily = 'Arial'"? That sounds quite painful... |
| 01:41 | <Lachy> | couldnt both .font as well as .fontFamily, .fontSize, etc. be provided like in CSSOM? |
| 01:42 | <Hixie> | the CSSOM needs to be optimised for reading |
| 01:42 | <Hixie> | this needs to be optimised for writing |
| 01:44 | <Lachy> | but if the author only wants to change one property, without affecting the others, then having to use the shorthand all the time would mean all properties would have to be explicitly specified, or they'd be inadvertenly reset to the defaults |
| 01:45 | <Lachy> | and changing, e.g. the font-weight requires specifying both the size and family at the same time |
| 01:45 | <Hixie> | that's the same argument as "what if you only want to change the Opacity component?" |
| 01:45 | <Hixie> | and the answer is, just implement the setter yourself |
| 01:45 | <Hixie> | instead of bloating the core API :-) |
| 01:46 | <Philip`> | Lachy: Out of lots of pages, I see approximately none setting .style.fontAnything |
| 01:47 | <Lachy> | Philip`, ok. |
| 01:47 | <Philip`> | (There's 5 setting style.font, 60 setting style.fontSize, 1465 setting style.display) |
| 01:47 | <Philip`> | (Not sure how many in total since my grep thing hasn't finished and it doesn't print any status information...) |
| 01:48 | <Philip`> | (*how many pages in total) |
| 01:49 | <Philip`> | (I really ought to compress all this data so it can be read from disk faster) |
| 01:49 | <Philip`> | (and also avoid having it spread over a hundred thousand separate files) |
| 01:53 | <othermaciej> | Hixie: my rough opinion after a bit of thought is that the current 'font' API is good as-is, but eventually canvas should be extended to have fonts as first-class objects too, so you can do things like change size or italics status or boldness level without having to do string pasting |
| 01:53 | <othermaciej> | Hixie: perhaps font objects could be added at the same time as path objects |
| 01:54 | <othermaciej> | Hixie: will try to send email along these lines soon |
| 01:54 | <Hixie> | k |
| 01:54 | <Hixie> | thanks |
| 02:31 | <Philip`> | Lachy: http://philip.html5.org/data/cssom-set-properties.txt shows the style properties that get set in inline scripts (or things that look quite like inline scripts) |
| 02:32 | <Philip`> | ('font' is still pretty far down) |
| 02:34 | <Philip`> | (Insert obvious warning about biased sample etc) |
| 02:46 | <Philip`> | Hixie: http://www.whatwg.org/specs/web-apps/current-work/status-documentation.html has typos in "Once a section is interoperably implemented, s quite stable and unlikely to change significantly" |
| 03:02 | <roc> | anyone staying at Wild Palms tonight? |
| 03:10 | <roc> | grrr |
| 03:11 | <othermaciej> | what's wrong? |
| 03:14 | <roc> | I posted to the wrong channel |
| 09:59 | <Hixie> | http://www.w3.org/MarkUp/2008/xhtml-basic-11-implementation.html is funny |
| 09:59 | <Hixie> | they only tested one implementation |
| 09:59 | <Hixie> | they only had 3 tests |
| 10:00 | <Hixie> | the implementation they tested was an internal build |
| 10:00 | <Hixie> | they modified the official test to run it |
| 10:03 | <othermaciej> | lol |
| 10:04 | <othermaciej> | someone should let them know that mobile profiling is dead |
| 13:21 | takkaria | boggles people |
| 14:13 | <hsivonen> | Hixie: Jeremy Keith pointed me to http://ufxtract.com/testsuite/ |
| 17:36 | <annevk> | back home, finally |
| 17:36 | <annevk> | GTA IV was in supply at the train station fortunately, so I'm all set |
| 17:36 | <annevk> | (on the other hand, the weather is quite nice around here...) |
| 20:07 | virtuelv | reads backlog across all of othermaciej's connects and disconnects |
| 20:07 | <virtuelv> | I almost died laughing reading that XHTML Basic test report |
| 20:08 | <Philip`> | virtuelv: http://krijnhoetmer.nl/irc-logs/whatwg/20080510 nicely hides the evidence of othermaciej's quite useless internet connection :-) |
| 20:09 | <virtuelv> | Philip`: so would my IRC client, if I just disabled that setting |
| 20:10 | <Philip`> | virtuelv: Does it apply that retroactively? |
| 20:11 | <virtuelv> | Philip`: i can't seem to find the setting directly in it |
| 20:11 | <virtuelv> | so I hardly think so |
| 20:12 | <virtuelv> | it doesn't, but I at least finally found the setting |
| 20:12 | <virtuelv> | nice bit is that XChat sets it on a per-channel basis |
| 20:35 | <jwalden> | postMessage has an early adopter: https://bugzilla.mozilla.org/show_bug.cgi?id=433183 |
| 20:35 | <jwalden> | and if I may say so, they are absolutely, utterly insane |
| 20:40 | <Philip`> | jwalden: Would it still be insane if they were closely following the feature's development and ensured their code always worked reliably with it? |
| 20:42 | <jwalden> | only slightly less so |
| 20:42 | <jwalden> | what fraction of a fraction of a percent of people would even be able to take advantage of it? |
| 20:43 | <jwalden> | once someone ships it makes sense |
| 20:43 | <jwalden> | but not before, except experimentally |
| 20:46 | <jwalden> | actually, I take that back |
| 20:46 | <jwalden> | implementations update to the spec at different speeds |
| 20:46 | <jwalden> | so even if you do follow it, you lose |
| 20:47 | <annevk> | opera had postMessage quite early |
| 20:47 | <annevk> | though it was exposed on document |
| 21:14 | <Philip`> | jwalden: I assume the three million FF3 beta testers would be able to take advantage of it, which is not insignificant |
| 21:16 | <jwalden> | they can't take that much more advantage of it than they can of whatever the supported method is in 2, and there's still the changing-implementation-rate problem |
| 21:21 | <othermaciej> | facebook?!?! |
| 21:27 | <jwalden> | so it seems |
| 21:29 | <BenMillard> | annevk, I'm back from Sight City with 14.5 pages of notes (A5 lined paper) |
| 21:30 | <BenMillard> | a ZoomText developer seemed keen to participate in HTML5 |
| 21:31 | <BenMillard> | other AT people seemed surprised that HTML5 even existed |
| 21:32 | <BenMillard> | they seemed generally positive about us wanting them to participate, though |
| 21:33 | <BenMillard> | the notes will eventually become a blog entry on Project Cerbera, somewhat like my entry for the HTMLWG F2F in Boston 2007 |
| 23:04 | Philip` | wonders if there's some way he can make Firefox 3 not totally fail to render borders around input boxes |
| 23:05 | <Philip`> | It makes textareas unpleasantly invisible :-( |