| 00:00 | Philip` | goes back to the great fun of trying to fix up the GPOS table, with its six zillion different types of data it can contain |
| 00:01 | Hixie | blog |
| 00:01 | <Hixie> | s |
| 00:01 | <Hixie> | my political correctness filter failed |
| 00:05 | <heycam> | weinig_, anyway i suspect that SVG wouldn't define things that "delay the load event" for the containing <iframe>, and that the iframe's load event would be dispatched as soon as the root <svg>'s load is dispatched |
| 00:05 | <Philip`> | "it can be far easier accomplished by changing the operational behavior of browsers such that they always send Referer" - it seems to quite (most?) often be proxies and firewalls (including software firewalls on end-users' machines) that strip Referer, so there's nothing the browser can do about it |
| 00:05 | <weinig_> | heycam: I think it is actually not that cut and dry |
| 00:05 | <heycam> | weinig_, @externalResourcesRequired can be used to delay the load event dispatched to the root <svg> |
| 00:05 | <Hixie> | heycam: is the root SVGLoad delayed until the font is loaded? |
| 00:05 | <heycam> | no? |
| 00:05 | <heycam> | Hixie, depends on eRR, i believe |
| 00:05 | <weinig_> | heycam: in SVG exteralResourcesRequired is used to delay |
| 00:05 | <weinig_> | indeed |
| 00:06 | <Hixie> | i can just hardcode that the SVG 'load' event is what we wait for if the resource isn't html |
| 00:06 | <heycam> | right, but it delays the load dispatched to the <svg> |
| 00:06 | <Hixie> | or whatever |
| 00:06 | <weinig_> | the root SVGLoad is only delayed if externalResourcesRequired=true |
| 00:06 | <heycam> | Hixie, that sounds like what i'd expect |
| 00:07 | <heycam> | i suppose the current text wouldn't work for that because SVG doesn't say anything about load events bubbling up to the window object |
| 00:07 | <heycam> | i guess dom 3 events will solve that? |
| 00:08 | <Hixie> | the current text doesn't work for html either |
| 00:08 | <heycam> | ah :) |
| 00:09 | <Hixie> | e.g. as written it implies you could fire a fake 'load' event and that'd be enough |
| 00:09 | <heycam> | i see |
| 00:09 | <Hixie> | crap, forgot my power cord. bbiab. |
| 00:24 | Hixie | uses someone else's |
| 00:34 | <Hixie> | woohoo, appcache feedback done |
| 00:35 | <Hixie> | shepazu: yt? |
| 00:50 | <campd> | Hixie: hrm |
| 00:50 | <campd> | Hixie: if I'm reading this right, a 404 of the manifest will cause an "error" to be fired to pending master entries |
| 00:50 | <Hixie> | right |
| 00:50 | <campd> | Hixie: but not a network/content-type/magic error |
| 00:50 | <campd> | is that intentional? |
| 00:50 | <Hixie> | network error should do, did i screw that up? |
| 00:50 | Hixie | looks |
| 00:51 | <campd> | misreading |
| 00:51 | <campd> | err, I think I misread |
| 00:51 | <campd> | yeah, misread |
| 00:51 | <campd> | never mind |
| 00:53 | <Hixie> | there are some cases where they get associated with an older cache, which come to think of it might break them... |
| 00:53 | <Hixie> | hm, crap, i should fix that |
| 00:57 | <campd> | :/ |
| 00:59 | <campd> | while you're at it, I saw a few typos |
| 00:59 | <campd> | want 'em here, or email? |
| 00:59 | <Hixie> | dump them here, i can fix them all at once |
| 00:59 | <Hixie> | since i'm editing it already |
| 00:59 | <Hixie> | thanks |
| 00:59 | <campd> | "optinally" |
| 01:02 | <campd> | hrm, lost the other one |
| 01:03 | <Hixie> | i found a "discrad" |
| 01:04 | <campd> | that's it |
| 01:05 | campd | will be spending tomorrow seeing how much of these changes we can fit in to 3.1 |
| 01:05 | <Hixie> | ok fixed all the typos and made the case of the manifest failing not cause the resources to get associated |
| 01:05 | <Hixie> | which would probably just break them and/pr poison the cache |
| 01:05 | <campd> | nod |
| 01:06 | <Hixie> | the new text is much better than the old text |
| 01:06 | <Hixie> | to start with it doesn't confuse cache and cache group all the time |
| 01:06 | <campd> | yeah |
| 01:10 | <campd> | Hixie: hrm, you think it's ok to violate no-store for different-origin entries? |
| 01:10 | <Hixie> | i dunno. right now the spec implies it (which is why i added the note pointing it out) |
| 01:10 | <campd> | yeah |
| 01:10 | <Hixie> | do you think we should drop them? seems weird to not cache things in the manifest. |
| 01:11 | <campd> | I agree |
| 01:11 | <Hixie> | ap said webkit keeps them, fwiw |
| 01:11 | <campd> | it makes sense to treat the manifest as superceding cache directives |
| 01:12 | <Hixie> | ok |
| 01:12 | <campd> | but it's kinda weird that evil.com's manifest can supercede normal.com's cache directives |
| 01:12 | <campd> | (it sounds worse when I call the first domain evil.com) |
| 01:13 | <campd> | (we implement it as currently specified, fwiw) |
| 01:14 | <Hixie> | yeah the cross-domain thing is weird, but, well, all it does is mean that if you go to evil.com you can end up with normal.com's sensitive information in your cache, right? |
| 01:14 | <campd> | yeah. |
| 01:15 | <Hixie> | hm, i just realised |
| 01:15 | <campd> | can't think of any real exploit/problem, just, y'know, weird. |
| 01:15 | <Hixie> | a site can use this to probe another site's filespace |
| 01:15 | <Hixie> | i guess you can already do that with iframes |
| 01:15 | <Hixie> | kinda |
| 02:44 | <Hixie> | annevk, i'm confused by the thread you're involved in (http://www.w3.org/mid/OFBC06637C.5D87622B-ON85257547.000D2A49-85257547.000DE226⊙lc) |
| 02:44 | <Hixie> | what's wrong with xhtml? |
| 02:48 | <Hixie> | wow |
| 02:48 | <Hixie> | adam posts http://www.adambarth.com/papers/2008/barth-jackson-mitchell-b.pdf |
| 02:49 | <Hixie> | a detailed paper with real data |
| 02:49 | <Hixie> | roy then posts http://www.w3.org/mid/3FD529DB-78D5-4561-AEAB-99F80BB26ADA⊙gc |
| 02:49 | <Hixie> | which makes assertions contradicting the data |
| 03:00 | <takkaria> | Roy does seem a tad idealistic |
| 03:02 | Philip` | finally convinces himself that he's probably handling the GPOS font table correctly, but sees that it apparently has precisely zero effect in the browsers he's testing |
| 03:02 | <Philip`> | (so I could just as well delete the entire table) |
| 03:06 | <Hixie> | what's it for? |
| 03:08 | <Philip`> | Mostly for fun, and otherwise for making subsets of fonts that won't take minutes to download over the web |
| 03:08 | <Hixie> | no i mean the GPOS table |
| 03:08 | <Philip`> | Oh |
| 03:08 | <Philip`> | "The Glyph Positioning table (GPOS) provides precise control over glyph placement for sophisticated text layout and rendering in each script and language system that a font supports." |
| 03:09 | <Hixie> | i... see |
| 03:09 | <Philip`> | e.g. like in http://www.microsoft.com/typography/otspec/fig4e.gif you can set it so that diacritics following an overhanging capital letter will be positioned downwards a bit |
| 03:10 | <Philip`> | or you can use it to make cursive scripts fit together better |
| 03:10 | <hdh> | is that the same as kerning? |
| 03:11 | <Hixie> | Philip`: wow, neat. and nobody does it? |
| 03:11 | <Hixie> | that's sad |
| 03:11 | <Philip`> | hdh: Kerning is much simpler - it's just an added horizontal offset between a pair of glyphs |
| 03:11 | <Philip`> | whereas this thing lets you use classes of glyphs, and more context than just pairs, and all sorts of other crazy stuff that I don't understand |
| 03:12 | <Philip`> | Hixie: I assume it's implemented somewhere, just not in the fonts and browsers and example strings that I'm using |
| 03:12 | <Hixie> | ah |
| 03:13 | <Philip`> | particularly since I'm miserly and using free fonts, of which there are very few that do advanced typography |
| 03:21 | <roc> | Philip`: what platform are you testing? Windows? |
| 03:21 | <roc> | We should support all the features that Uniscribe does, since we just use Uniscribe for this |
| 03:23 | <Philip`> | roc: Linux |
| 03:23 | Philip` | should probably try it on Windows some time |
| 03:23 | <Philip`> | but after going to bed :-) |
| 03:25 | <roc> | well then, on Linux we support everything Pango supports |
| 03:25 | <Hixie> | well bummer, my next task is to work out how UAs handle content-type for <Script> |
| 03:26 | <Hixie> | what i expect to find is taht they ignore the type and just look for a charset... |
| 03:26 | <Hixie> | even if the type doesn't allow charset |
| 03:28 | <Hixie> | hey wow, i have some tests that are polyglots! |
| 03:29 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/script/008-test-plain-none.txt |
| 04:00 | <Hixie> | if someone can work out a way of making my a polyglot js file that is valid JS in both UTF-8 and UTF-16, that would be awesome |
| 04:00 | <Hixie> | even if all it did was a() in UTF-8 and b() in UTF-16 |
| 04:01 | <Hixie> | wow |
| 04:01 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/script/011-demo.html |
| 04:01 | <Hixie> | IE8 ignores the Content-Type header altogether |
| 04:02 | <Hixie> | even for ;charset= processing |
| 04:02 | <Hixie> | that's crazy! |
| 04:02 | <Hixie> | nobody else does |
| 04:02 | <Hixie> | every other browser looks for ;charset="" |
| 04:02 | <Hixie> | did i do something wrong maybe? |
| 06:21 | <othermaciej> | I admire abarth's patience |
| 06:23 | <MikeSmith> | I admire abarth in general |
| 06:23 | <MikeSmith> | but I do hope his patience holds out |
| 06:23 | <othermaciej> | was just reading the Origin thread |
| 06:23 | <MikeSmith> | yeah, I figured that's what you were referring to |
| 06:24 | <jruderman> | is this the same thread where someone claimed CSRF is just a browser bug? |
| 06:24 | <MikeSmith> | jruderman: the someone being Roy Fielding |
| 06:25 | <MikeSmith> | though I don't think he worded it that way exactly |
| 06:27 | <MikeSmith> | "CSRF is not a security issue for the Web... If browsers create a security issue because they allow scripts to automatically direct requests with stored security credentials onto third-party sites, without any user intervention/configuration, then the obvious fix is within the browser." |
| 06:27 | <MikeSmith> | is what he said |
| 06:27 | <MikeSmith> | http://lists.w3.org/Archives/Public/ietf-http-wg/2009JanMar/0037.html |
| 06:31 | <othermaciej> | I guess I should subscribe to ietf-http-wg |
| 06:37 | jruderman | realizes that Roy Fielding is one of the authors of the HTTP spec |
| 06:45 | <othermaciej> | jruderman: also the author of the famous REST architecture paper |
| 06:45 | <othermaciej> | jruderman: (the paper that coined the term -- I guess it was his thesis) |
| 06:52 | <Hixie> | othermaciej: abarth is a god |
| 06:52 | <Hixie> | (that is my professional opinion) |
| 06:53 | <Hixie> | however if he applies for any jobs you should NOT hire him. _I_ want him. :-P |
| 06:54 | <Hixie> | on another note, does anyone have any insight into who IE behaves so very differently to the other browsers on http://www.hixie.ch/tests/adhoc/html/script/011-demo.html ? |
| 06:55 | <Hixie> | (it seems to ignore the ;charset=... part of the MIME type, unlike other UAs) |
| 06:57 | <othermaciej> | Hixie: I don't generally actively poach from other browser vendors but if he applies, I make no promises |
| 08:14 | <Hixie> | othermaciej: :-) |
| 08:15 | <Hixie> | ok seriously, what's with IE |
| 08:15 | <Hixie> | I don't get it |
| 08:15 | <Hixie> | does it not support setting the character encoding on the fly at all? |
| 08:15 | <Hixie> | Philip`: do you have data on mime types that <script src=""> files are sent with? |
| 08:19 | <Hixie> | the web seems to suggest that IE really doesn't support this, wow |
| 08:25 | <jruderman> | http://www.whitehouse.gov/includes/functions.js.aspx is text/html |
| 08:25 | <jruderman> | that file also contains some head-spinners, like |
| 08:25 | <jruderman> | if ( isLeapYear(y)) MonthArray[1] = eval(eval(MonthArray[1]) + 1); |
| 08:27 | <Hixie> | does MonthArray contain strings or numbers? |
| 08:27 | <jruderman> | well, uh |
| 08:27 | <jruderman> | strings, up until that line |
| 08:27 | <jruderman> | Months = "31/!/28/!/31/!/30/!/31/!/30/!/31/!/31/!/30/!/31/!/30/!/31"; |
| 08:27 | <jruderman> | MonthArray = Months.split("/!/"); |
| 08:28 | <Hixie> | wow, that question led to more pain than i expected |
| 08:28 | <jruderman> | after that line, it might contain one number |
| 08:28 | <Hixie> | I wonder what led to that code |
| 08:29 | <jruderman> | hehe |
| 08:36 | <annevk> | oh lol, Roy was pwned, to to love abarth |
| 08:36 | <annevk> | got to* |
| 08:37 | annevk | forgets a lot of words in his IRC remarks for some reason |
| 08:39 | jgraham | is loving the WHATWG email that predicts online shops and auction sites within the next five years |
| 08:39 | <jruderman> | url? |
| 08:41 | <jgraham> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-January/018329.html |
| 08:42 | <jgraham> | (to be fair it is basically a sensible email so it is probably just a non-native speaker) |
| 08:42 | <annevk> | http://twitter.com/diveintomark/status/1140911457 lol |
| 08:54 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/script/011-demo.html is a ridiculously network-intensive test. |
| 08:57 | <hsivonen> | "When you’re deciding what language to write your web app in, the big advantage of Flex is that it actually exists" |
| 08:58 | <hsivonen> | http://otherfancystuff.blogspot.com/2009/01/html5-vs-flex-for-rias-ignite-style.html |
| 08:59 | <Hixie> | oh hey, my network-abusing neighbour is back |
| 08:59 | <annevk> | webben_, you should write tutorials man; you write many long good e-mails, but they all get lost in archives unless you collect them somehow (you could put stuff on blog.whatwg.org or wiki.whatwg.org) |
| 08:59 | <Hixie> | good thing i've been logging every packet sent out from my network |
| 08:59 | Hixie | delves into the packet logs to see who this is |
| 09:00 | Hixie | got one of his other neighbours to stop using his network very suddenly by redirecting all his web traffic to a short page noting that he knew the person's name... right bang in the middle of that person browsing cross-dresser support forums. |
| 09:03 | <jruderman> | support as in technical support or support as in moral support? |
| 09:04 | <Hixie> | both, actually |
| 09:04 | <Hixie> | he was shopping for fake breast inserts |
| 09:04 | <Hixie> | and browsing sites about people's first public outting as their "other" personality |
| 09:04 | <Hixie> | i should point out i have nothing against cross-dressers |
| 09:04 | <Hixie> | and didn't even mention this in the message |
| 09:05 | <Hixie> | but they stopped so suddenly, i think they may have been freaked out a little |
| 09:05 | <jruderman> | you just have something against other people using your network? |
| 09:05 | <jruderman> | or using it too heavily? |
| 09:05 | <Hixie> | (i actually didn't realise what they were doing when i put up the message, only found out later when checking the logs more carefully) |
| 09:05 | <Hixie> | just too heavily |
| 09:05 | <hsivonen> | it seems that our FAQ doesn't explain why HTML5 and XHTML2 are separate |
| 09:05 | <webben_> | annevk: Thanks :) And yeah, I do need some sort of useful-info-in-email-preservation approach, and not just for html5-related things. ;) |
| 09:05 | <annevk> | Hixie, dunno about the TAG thread, though it seems Noah thinks there is more legacy in tools than in content? |
| 09:05 | <Hixie> | jruderman: if i notice they are using it, then they're using it too heavily. then i start looking. :-) |
| 09:06 | <Hixie> | jruderman: basically, i provide it as a public service for people who want to check their e-mail while at a party, or who want to contact comcast to set up their own network |
| 09:06 | <Hixie> | jruderman: if they're browsing porn... or downloading HD movies... they can get their own network |
| 09:07 | <Hixie> | annevk: dunno |
| 09:08 | annevk | leaves it |
| 09:27 | <hsivonen> | does Prince support gzip? |
| 09:27 | <hsivonen> | (in HTTP) |
| 09:59 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/script/011-demo.html shows firefox has really weird content-type parsing |
| 09:59 | <Hixie> | it's not the simple algorithm in the spec (or in webkit) but it's not the spec's proper algorithm either |
| 10:00 | <Hixie> | opera is per spec, it seems |
| 10:01 | <annevk> | HTTP? |
| 10:02 | <Hixie> | yeah |
| 10:03 | <annevk> | some browser has to care about the standards I suppose 8-) |
| 10:03 | <hsivonen> | what happened to Mr. This Week? |
| 10:05 | annevk | asks on twitter |
| 10:09 | <Philip`> | Hixie: I have http://canvex.lazyilluminati.com/misc/stats/scripttypes2.html which I *think* is script content-types, but it's from so long ago that I can't quite remember |
| 10:15 | <Hixie> | he's fine |
| 10:15 | <Hixie> | just been taking a break because i haven't done much |
| 10:15 | <Hixie> | he said he'd probably start again next week or so |
| 10:15 | <Hixie> | Philip`: looking... |
| 10:15 | <Hixie> | Philip`: awesome, thanks |
| 10:16 | <annevk> | maybe I should update html5-diff today so that we can publish sooner rather than later |
| 10:17 | <annevk> | though digging through SVN back to June last year is not so attractive |
| 11:00 | <Philip`> | annevk: s/incorperated/incorporated/ in html5-diff |
| 11:03 | <annevk> | ta |
| 11:08 | Philip` | tries Firefox nightly (3.2a1pre) on Windows, and finds that it's incredibly slow at rendering his font test page |
| 11:08 | <Philip`> | like it takes about twenty seconds to load all the fonts |
| 11:08 | <Philip`> | using all my CPU power while it's doing so |
| 11:15 | <roc> | URL? |
| 11:16 | <Philip`> | http://192.168.2.3/... |
| 11:16 | <roc> | pff |
| 11:17 | Philip` | will try uploading it somewhere in a few minutes |
| 11:30 | <hsivonen> | a friend of mine got bitten by someone actually implementing RFC 3023 :-( |
| 11:34 | <annevk> | did he have to go to the hospital? |
| 11:36 | <Hixie> | do they have shots for that? |
| 11:37 | <annevk> | yeah, HTML5 content sniffing :p |
| 11:37 | <hsivonen> | I gather there are t-shirts |
| 11:38 | <Hixie> | what's wrong with 3023 again? just the text/xml charset nonsense? |
| 11:38 | <hsivonen> | Hixie: yes |
| 11:38 | <Hixie> | ok |
| 11:41 | <Hixie> | hsivonen: i've been forgetting to mark recent changes as applying to conformance checkers, sorry |
| 11:42 | <hsivonen> | Hixie: ok. I'll review the changes more closely |
| 11:44 | <Hixie> | in particular, 2675 and 2697 probably affect you. |
| 11:51 | <Philip`> | It's strange how much less correct code looks after you wait until the next day and then look at it again |
| 12:19 | annevk | notices an instance of "(blame: hs)" rather than "(credit: hs)" and laughs |
| 12:38 | <annevk> | jgraham, html5-diff is a WD |
| 12:39 | Philip` | finds a second bug in Font::TTF, which is pretty good going so far |
| 12:39 | Philip` | supposes that's an advantage of using a library that's more than a decade old |
| 12:42 | <hsivonen> | does Opera 10 support the new name for <eventsource>? |
| 12:43 | <jgraham> | annevk, karlcow: At what point do you have to decide if a WD will become a TR or a Note? |
| 12:43 | <hsivonen> | jgraham: wouldn't it be non-sensical to make it a REC? |
| 12:44 | <hsivonen> | jgraham: assuming you meant REC. (Notes are under /TR/, too.) |
| 12:44 | <karlcow> | jgraham: any point. |
| 12:45 | <jgraham> | hsivonen: I was under the impression that some people had suggested making the Markup Spec normative |
| 12:46 | <hsivonen> | jgraham: I though you meant the diff document |
| 12:46 | <jgraham> | s/Markup Spec/HTML 5 The Markup Language" |
| 12:46 | <karlcow> | TR Process is a publication process architectured around a technology and its maturation. What people fail to understand often, that there is no exit criteria from one step to the other, but *entrance criteria* |
| 12:48 | <hsivonen> | karlcow: is there any way in the Process for marking a WD dead if it fall out of charter or the owning group decides against advancing it? |
| 12:49 | <hsivonen> | karlcow: for example, is there a way to flag the Web Forms 2.0 dead as a separate document? |
| 12:50 | <Philip`> | Wow! I found a situation where the GPOS table actually does something visible |
| 12:50 | <Philip`> | Specifically, spacing of "VA" and "To" in Libertine and Deja Vu fonts on Windows only |
| 12:50 | <annevk> | jgraham, you usually state the intent upfront |
| 12:51 | <Philip`> | (and only in Firefox, not in IE) |
| 12:51 | <hsivonen> | annevk: yeah, there's something fishy going on if the intent of REC vs. Note isn't stated up front |
| 12:51 | <annevk> | hsivonen, we should simply redirect the web-forms-2 shortname to html5 when we publish a new draft |
| 12:52 | <hsivonen> | Philip`: how does that work regarding cross-platform line breaking in OOo? |
| 12:52 | <annevk> | hsivonen, I suppose we should e-mail someone about that to make sure it happens though |
| 12:57 | <Philip`> | hsivonen: I would imagine such things are inconsistent across platforms |
| 14:01 | <karlcow> | hsivonen: draft ideas are never dead :) luckily enough :) |
| 14:03 | <karlcow> | hsivonen: nothing fishy. Strange way of conceiving the evolution of ideas and life of things. It is a bit like if you were saying "wow this guy is not saying up front if he wants to become the director of the company or a developer in this company". |
| 14:04 | <karlcow> | Things evolve, change. It is not binary and deterministic. |
| 14:11 | <karlcow> | hsivonen: >For years, HTML4 validation wasn't improved at all. Validator.nu, Validome and Relaxed have now improved on things but are still seen as illegitimate compared to DTD-based HTML4 validation. |
| 14:11 | <karlcow> | this is quite ironic |
| 14:12 | <Lachy> | karlcow, in what way is that ironic? |
| 14:12 | <karlcow> | for years, w3c has asked people for help, but people prefer to create their own competing services more than improving the one in place. :) The way you tell the story gives an interesting twist to what is happening |
| 14:40 | <hsivonen> | karlcow: I think we've had this discussion before. Technical improvements required a total engine swap and coming into a project from the outside advocating a total engine swap wouldn't have been socially feasible without first demonstrating technical merit elsewhere. |
| 14:46 | <karlcow> | hsivonen: which is a complete different take on your email. |
| 15:02 | <hsivonen> | karlcow: part of the problem were the people who held the line "validation has a clear and precise meaning. it's parsing without DTD-based errors. nothing more, nothing less" |
| 15:03 | <hsivonen> | karlcow: (in the community around the W3C validator) |
| 15:03 | <karlcow> | hsivonen: your pushing the discussion elsewhere |
| 15:04 | <hsivonen> | karlcow: I thought I pushed it back to normative schemas (DTDs in the case of HTML4) |
| 15:04 | <karlcow> | the same will happen to your validator if the community decide to make it the emblem of html 5 validation. Unfortunately not your choice, power of the crowd |
| 15:10 | <annevk> | idea: create a XHTML2 versus HTML5 wiki page discussing XHTML2 chapter by chapter, saying how HTML5 handles the particular feature and what the rationale was doing it that way |
| 15:10 | annevk | hopes to work on that at some point |
| 15:12 | <zcorpan> | or the other way around? |
| 15:13 | <annevk> | then I wouldn't find out what XHTML2 does and HTML5 doesn't |
| 15:13 | <zcorpan> | true |
| 15:13 | <annevk> | but someone could create a HTML5 versus XHTML2 page, sure :) |
| 15:15 | <hsivonen> | I'd rather see the effort spent of improving Lachy's authoring guide, for example |
| 15:16 | <annevk> | hsivonen, I don't think I'm a good tutorial author |
| 15:17 | <annevk> | writing books and such is best done before being an expert, according to markp anyway |
| 15:18 | <hsivonen> | I wonder if markp considers himself to past the point of being able to write Dive into HTML5. |
| 15:19 | <karlcow> | annevk: yes it is a nice idea |
| 15:19 | <Lachy> | I should have the time to get back to working on the authoing guide soon |
| 15:19 | <annevk> | also, to be honest, I think DanC's approach is better |
| 15:19 | <karlcow> | hsivonen: it is difficult to work on Lachy tutorial |
| 15:19 | <annevk> | what he outlined anyway |
| 15:19 | <karlcow> | agreed with annevk |
| 15:19 | <hsivonen> | karlcow: why? |
| 15:19 | <Lachy> | what's DanC's apparoch? |
| 15:19 | <hsivonen> | where's DanC's outline? |
| 15:20 | <karlcow> | because it is not a tutorial. |
| 15:20 | annevk | looks |
| 15:20 | <karlcow> | it is not about explaining use cases and how to solve them |
| 15:20 | <karlcow> | with html5 |
| 15:20 | <annevk> | http://lists.w3.org/Archives/Public/public-html/2009Jan/att-0274/html-writing-ideas.html |
| 15:21 | <karlcow> | I was tempted to start from scratch but have been holding it back by the friction that it would create with Lachy |
| 15:21 | <annevk> | http://lists.w3.org/Archives/Public/public-html/2009Jan/0274.html is the e-mail that goes with it |
| 15:21 | <karlcow> | s/by/because of |
| 15:21 | <Lachy> | how is that outline significantly different from what my authoing guide will be? |
| 15:22 | <annevk> | see the e-mail |
| 15:23 | <karlcow> | simple bits did that in the past |
| 15:23 | <karlcow> | http://www.simplebits.com/publications/solutions/toc/ |
| 15:23 | <karlcow> | and it was quite cool |
| 15:23 | <karlcow> | unfortunately he removed the content when the book was ready |
| 15:23 | <karlcow> | more than saying this is a ul and how to use it |
| 15:23 | <karlcow> | it was on the form |
| 15:23 | <karlcow> | How do I do lists in documents |
| 15:24 | <karlcow> | what is a quote and how do I use it |
| 15:24 | <karlcow> | etc |
| 15:24 | <Lachy> | I'm still not getting what the difference is. Perhaps my view of what the guide will be doesn't quite match up with what people think I doing withit |
| 15:25 | <hsivonen> | I wish a guide didn't try to dumb things down and started boldly with the concept of a document tree |
| 15:25 | <karlcow> | Lachy: the way I look at your document for now is what the Markup Spec is doing somehow. An inventory of vocabulary |
| 15:25 | <hsivonen> | hmm. Lachy's doc does mention the tree early on, but doesn't illustrate it in detail |
| 15:27 | <zcorpan> | hsivonen: i seem to remember doing just fine making pages without knowing what the DOM was |
| 15:27 | <Lachy> | karlcow, ok. I think I need to create a clearer outline so people can put what's currently in the document, in perspective of everything it will contain |
| 15:27 | <zcorpan> | at least until i started to do scripting |
| 15:28 | <Philip`> | You can still do quite a bit of scripting without knowing anything about a document tree |
| 15:28 | <zcorpan> | indeed |
| 15:28 | <karlcow> | Lachy: sincerely I would dump everything about the vocabulary inventory |
| 15:28 | <karlcow> | this should be the last thing |
| 15:28 | <karlcow> | not the first one |
| 15:28 | <Philip`> | document.getElementById('foo').setAttribute('style', 'text-decoration: blink') and all that stuff |
| 15:28 | <zcorpan> | i also remember that the DOM was just confusing and irrelevant before learning it |
| 15:28 | <Philip`> | Also you can do a lot of CSS without knowing about trees, just by referring to everything directly with element names and class names |
| 15:30 | <Lachy> | karlcow, having a description of the vocabulary is the most important part, IMHO |
| 15:31 | <Lachy> | we shouldn't need to refer authors to the spec just to get the vocabulary |
| 15:31 | <annevk> | hsivonen, maybe if you have a CS background "tree" makes sense, if you're still in high school you just want to know how to get a webpage online :) |
| 15:32 | <karlcow> | well exactly why I have not been working on it ;) I indeed prefer the document proposed by danc |
| 15:32 | <Lachy> | DanC's outline clearly incldues a whole section covering the vocabulary. I don't see the difference |
| 15:32 | <annevk> | have you read his e-mail? |
| 15:33 | <Lachy> | yes |
| 15:33 | <annevk> | to me it indicates he's planning on writing separate tutorials on how to do various things, such as putting a video online |
| 15:34 | <annevk> | i think that's vastly different from a spec that describes the HTML5 vocabulary |
| 15:34 | <karlcow> | yes |
| 15:35 | <Lachy> | describing the vocabulary is just part of the guide. I also intended to describe use cases like that |
| 15:35 | <Lachy> | the way I see it, it's all related and having DanC work on a separate document, rather than just contributing to the one we already have is pointless |
| 15:37 | <annevk> | you still can't see how your approach is different? having everything in the same doc is vastly different from having lots of little tutorials |
| 15:39 | <annevk> | and I can't at all see why you'd consider it pointless, if he writes something and it's good, you can always borrow it if the WG decides to make an authoring guide |
| 15:40 | <Philip`> | It seems logistically easier to have separate people working on separate documents (as long as they're not duplicating each others' work), and later the documents that end up being good can be focussed on and maybe made more consistent and merged together |
| 15:40 | <Lachy> | I don't see why it should be done in a separate document. He's welcome to write stuff though |
| 15:41 | <Lachy> | perhaps we can just merge them later though, once I actually see a more concrete proposal |
| 15:42 | <annevk> | the reason it should be in a separate document is because that makes it more approachable to authors |
| 15:42 | <annevk> | concretely, "HOWTO put video on the Web" is more approachable than "HTML5 Authoring Guide" (and also a lot smaller and readable within a few minutes) |
| 15:44 | <Lachy> | alright, I suppose we can make both documents compliment each other |
| 15:44 | <hsivonen> | will the video be Ogg? |
| 15:49 | <Lachy> | ok, if DanC's document will follow the cookbook approach, which isn't impression I get from either his outline or his email, then I can accept that |
| 15:54 | <annevk> | http://groups.google.com/group/demon.homepages.authoring/browse_thread/thread/ccdef4ff7b26c4ca |
| 16:01 | <Lachy> | LOL. It's been a while since I've read that :-) |
| 16:07 | <Lachy> | the <cowpats> element would be quite useful for some people :-) |
| 16:09 | <Dashiva> | How about <cowpath> to enclose content you feel should be made part of the next spec iteration? :) |
| 16:17 | <Philip`> | Dashiva: But what markup would you use to indicate that you feel <cowpath> should be made part of the next spec iteration? |
| 16:17 | <Dashiva> | <div class="cowpath"> obviously |
| 16:38 | <Philip`> | Hmm, the Libertine font has a section saying that any glyph in the range 1945..1964 should (in certain situations) be replaced by the glyph with an index 20 places earlier. But I'm removing arbitrary glyphs and reordering them all, so I can't use the same delta for each, so this seems pretty much impossible to handle :-( |
| 17:53 | <annevk> | http://wiki.whatwg.org/wiki/XHTML2_versus_HTML5 has a start of my idea |
| 17:53 | <gsnedders> | annevk has an idea? Oh shit. |
| 17:58 | <Philip`> | Someone should make an 'XHTML2 versus HTML5' version of http://coop.sigkill.com/~preed/people.mozilla.com/~preed/bloggity-blog-blog/vcs-shootout3.jpg |
| 18:42 | <annevk> | hmm, XHTML2 has a layout="" attribute |
| 18:44 | <Philip`> | <table layout="layout">? |
| 18:44 | <annevk> | http://www.w3.org/MarkUp/2009/ED-xhtml-modularization2-20090123/mod-core.html |
| 18:45 | <annevk> | hmm http://www.w3.org/MarkUp/2009/ED-xhtml-modularization2-20090123/mod-hyperAttributes.html in XHTML2 defines some user agent processing in a strikingly similar manner to HTML5 |
| 18:46 | <gsnedders> | So then with @layout the element hasLayout? |
| 18:53 | <Lachy> | the layout attribute is their method of making every element behave like <pre> |
| 18:53 | <gsnedders> | Does xml:space do that? |
| 18:53 | <Lachy> | no |
| 18:54 | <Philip`> | The layout attribute doesn't unless you also use xml:space, as far as I can tell |
| 18:54 | <Philip`> | (because otherwise some of the space will get normalised after coming out of the XML parser) |
| 18:55 | <annevk> | XHTML2 is such a poor spec |
| 18:56 | <Lachy> | from XML spec: "The value "default" [for xml:space] signals that applications' default white-space processing modes are acceptable for this element" |
| 18:57 | <Lachy> | it doesn't seem to say what the default white-space processing mode is. It seems to be application dependent |
| 18:57 | <Lachy> | AFAICT, the xml:space attribute has no effect on the parser |
| 18:58 | <annevk> | the table stuff is just a copy of HTML4, including the "A graphical user agent might render this as:" with a screen shot from an age old user agent |
| 18:58 | <Lachy> | heh |
| 19:02 | <annevk> | can't believe people are using XHTML2 in arguments, even hypothetical ones |
| 19:04 | <Lachy> | who is? |
| 19:04 | <Lachy> | do you mean on one of the mailing lists, or in general? |
| 19:05 | <annevk> | general |
| 19:06 | <Lachy> | ok. that explains why I didn't see any such arguments on the list recently |
| 19:06 | <annevk> | anyway, it still seems semi-useful to explore it for markup ideas and document that somehow |
| 19:06 | <annevk> | also means I have some place to point to when anyone asks |
| 19:13 | <annevk> | http://twitter.com/diveintomark/status/1142232859 :p |
| 19:35 | <Philip`> | hsivonen: Is it undesirable for me to have validator.nu say "Internal Error: Oops. That was not supposed to happen. A bug manifested itself in the application internals. Unable to continue. Sorry. The admin was notified."? |
| 20:20 | <annevk> | http://twitter.com/robburns1 |
| 20:22 | <Lachy> | who does he think is only looking at XHTML2 for the first time? |
| 20:22 | <gsnedders> | Lachy: annevk |
| 20:23 | <annevk> | he obviously never heard of Google |
| 20:23 | <annevk> | http://www.google.com/search?q=site:annevankesteren.nl+xhtml2 |
| 20:31 | <annevk> | WHATWG blog still uses XHTML 1.0 Transitional |
| 20:31 | <annevk> | :) |
| 20:53 | gsnedders | wonders whether to tweet how he feels |
| 20:57 | <Philip`> | Is it wise to constrain your feelings to 140 characters? |
| 20:57 | <jcranmer> | Philip`: I feel... complex |
| 20:57 | <jcranmer> | :-) |
| 21:00 | <gsnedders> | Philip`: Oh, I can easily live within that limit. |
| 21:01 | <gsnedders> | Philip`: The question is how do I word it vague enough for nobody to get who it is about but not have it so vague that it's meaning is obscured |
| 21:04 | <krijnh> | -_- |
| 21:04 | <gsnedders> | :P |
| 21:04 | <gsnedders> | (Also, it probably isn't as bad as that makes it sound.) |
| 21:06 | <Philip`> | gsnedders: If you don't actually want anybody to know what you're feeling, maybe it's easier to not express it at all :-) |
| 21:09 | <Philip`> | Hooray, now Firefox renders Arabic in a subsetted Deja Vu Sans more correctly |
| 21:10 | Philip` | probably has millions of bugs in this code, but the output looks alright in his tests so he's just going to leave it for now |
| 21:15 | <Lachy> | gsnedders, your last tweet was quite vague. |
| 21:15 | <gsnedders> | Lachy: Quote from Ada or Ardor |
| 21:16 | <Lachy> | who are they? |
| 21:16 | <gsnedders> | I would've tweeted the next sentence too, but that would've made it too long |
| 21:16 | <gsnedders> | s/Ada or Ardor/"Ada or Ardor" |
| 21:16 | <gsnedders> | Novel, by Vladimir Nabokov |
| 21:16 | <Lachy> | ok |
| 21:17 | <gsnedders> | The next sentence is, off hand, "But a butterfly in the Park, an orchid in a shop window, would revive everything with a dazzling inward shock of despair." |
| 21:17 | <Lachy> | well, I'm waiting in suspence to see if you can top the vagueness with your next tweet |
| 21:17 | <gsnedders> | Lachy: Vagueness is not the aim of the next tweet. Not revealing who it is about is the aim. :) |
| 21:20 | <Lachy> | is it about that same girl you tweeted about yesterday? |
| 21:20 | <gsnedders> | What did I tweet yesterday? |
| 21:20 | <gsnedders> | :) |
| 21:20 | <Lachy> | http://twitter.com/gsnedders/status/1137360266 |
| 21:20 | <gsnedders> | That was two days ago. |
| 21:21 | <gsnedders> | No, nothing to do with her. |
| 21:21 | <gsnedders> | http://twitter.com/gsnedders/status/1137514005 is about her too. |
| 21:21 | <Lachy> | it says 8:36 yesterday |
| 21:21 | <gsnedders> | Nothing more than that needs to be said |
| 21:21 | <gsnedders> | Lachy: You and your wacky timezones |
| 21:21 | <Lachy> | yeah, I figured that much |
| 21:21 | <Lachy> | My timezone is only about an hour or so different from yours |
| 21:22 | <Lachy> | must be twitter using US time |
| 21:22 | <annevk> | but twitter.com probably isn't |
| 21:22 | <gsnedders> | It was around 48 hours ago |
| 21:22 | <gsnedders> | "You want the honest answer? I still want you. Time changes nothing." — That's the draft I have of the next tweet, minus the last sentence. I fear the last sentence will give it away. |
| 21:23 | <krijnh> | You can say it here.. I'll take the logs offline for you ;) |
| 21:23 | <gsnedders> | :P |
| 21:23 | <gsnedders> | Whatever you say krijnh, whatever you say :) |
| 21:24 | <Philip`> | gsnedders: The "@Hixie:" would give it away too |
| 21:24 | <gsnedders> | :) |
| 21:25 | <gsnedders> | This person doesn't have Twitter. |
| 21:25 | <krijnh> | You can use [off] |
| 21:25 | <gsnedders> | That works in here too? |
| 21:25 | <gsnedders> | [off] foobar |
| 21:25 | <krijnh> | Oh, it doesn't |
| 21:25 | <gsnedders> | I thought #webapps was special cased |
| 21:25 | <gsnedders> | And the logs would tend to confirm that |
| 21:26 | <Hixie> | you can't hide your feelings here! we're radically open and transparent! |
| 21:26 | <Hixie> | sweet, danc is going to take some of my work off my hands |
| 21:26 | <gsnedders> | Okay okay okay… The truth! I'm not gay. |
| 21:27 | <Lachy> | gsnedders, we know that much |
| 21:27 | <krijnh> | We do? |
| 21:27 | <gsnedders> | krijnh: See tweets linked a few minutes ago |
| 21:27 | <Lachy> | well, the way he obsesses over girls he thinks he can never have is a big giveaway |
| 21:28 | <krijnh> | gsnedders: sorry, I don't care enough about you being gay or not ;) |
| 21:28 | <gsnedders> | obsesses is too strong of a word. |
| 21:28 | <Lachy> | ok, if you say so |
| 21:28 | <gsnedders> | :D |
| 21:29 | <Hixie> | maybe he only obsesses over the girls because he's already getting the boys |
| 21:29 | <Lachy> | is she someone you are freinds with, or just someone from school you like, who hangs out with different people? |
| 21:29 | <gsnedders> | Well, the friends I spend most of my time with are all girls and I fit in far too well so I do get called gay far too often |
| 21:30 | <gsnedders> | Lachy: no comment. |
| 21:30 | gsnedders | will not be cross-examined about this here |
| 21:30 | <gsnedders> | (At least insofar as who the draft is about) |
| 21:30 | <krijnh> | One disadvantage of bringing it up |
| 21:31 | <gsnedders> | krijnh: I'm used to a group of very inquisitive girls questioning me about such things. I'm sure I can manage. |
| 21:32 | <Lachy> | gsnedders, that's what I thought when I was your age. But somehow, people used to get all sorts of information from me I didn't want to reveal |
| 21:33 | <gsnedders> | Lachy: I can be very quiet if I want to be :) |
| 21:33 | <Philip`> | gsnedders: Choosing to be quiet is an act that reveals information too |
| 21:34 | <gsnedders> | Philip`: No, it isn't my sister. |
| 21:34 | <gsnedders> | :P |
| 21:34 | <takkaria> | choosing to be quier is not an act per se |
| 21:34 | <gsnedders> | "If you choose not to decide/You still have made a choice" |
| 21:35 | <Philip`> | takkaria: How is it not an act? |
| 21:36 | <takkaria> | an act generally has some positive element |
| 21:36 | <takkaria> | e.g. it is an act rather than choosing to be inactive |
| 21:37 | <gsnedders> | Lachy: As for not giving stuff away, I haven't yet given away the gender of who it is about |
| 21:37 | <Philip`> | It's like if you're sending secret encrypted instructions to an undercover agent in an enemy country and then suddenly stop sending anything at all, and that act will give information to an eavesdropper that might let them deduce the agent is no longer active |
| 21:39 | <annevk> | once gsnedders does HTML2RFC I can attempt to register about: |
| 21:39 | <gsnedders> | annevk: I refuse. |
| 21:39 | <gsnedders> | I have henceforth said I will not write html2rfc, and I have no intent on changing that POV. |
| 21:40 | <annevk> | :/ |
| 21:40 | <Philip`> | How about html2rfcxml? |
| 21:40 | <Hixie> | annevk: i can send you my script |
| 21:40 | <Lachy> | gsnedders, it's either a girl, or a boy you refer to as "her". My guess is that it's not the latter. |
| 21:40 | <Hixie> | annevk: but you'll have to hack it |
| 21:41 | <gsnedders> | Lachy: When did I refer to them as "her"? |
| 21:41 | <Lachy> | "I want to be with her so badly, but I know she'll never have me." |
| 21:42 | <Lachy> | or are you talking about someone else now? |
| 21:42 | <gsnedders> | Lachy: I've already said it had nothing to do with her. |
| 21:42 | <Lachy> | oh |
| 21:42 | <Lachy> | ok |
| 21:42 | <gsnedders> | (She does, inadvertently, know who this is about.) |
| 21:42 | <gsnedders> | (And would probably get it from the vague draft I quoted above.) |
| 21:42 | <annevk> | Hixie, I'm not too great with Perl |
| 21:43 | <gsnedders> | annevk: Go learn. |
| 21:43 | <Hixie> | annevk: who is |
| 21:43 | <Hixie> | gsnedders: thanks for the templates btw |
| 21:43 | <gsnedders> | Hixie: np |
| 21:45 | <Hixie> | annevk: http://damowmow.com/temp/html-to-rfc.txt |
| 21:45 | <Hixie> | it's more generic that i remember |
| 21:46 | <Hixie> | basically you give it the ID of an <h4>, and it extracts that section and its subsections |
| 21:46 | <gsnedders> | Every time I look at Perl I just think "oh god" |
| 21:47 | <Lachy> | Hixie, what would an I-D to register the about: URI scheme need to say? |
| 21:47 | <Hixie> | "about:blank returns the empty string" |
| 21:47 | <Hixie> | i think that might be it! |
| 21:47 | <Hixie> | i guess it would have to define valid values of about:... |
| 21:47 | <Lachy> | heh |
| 21:48 | <Hixie> | and what they represent |
| 21:48 | <Hixie> | some history |
| 21:48 | <Hixie> | some mention of the security implications of the origin of about:Blank |
| 21:48 | <Hixie> | and how it must be done like html5 says |
| 21:48 | <Hixie> | or something |
| 21:48 | <gsnedders> | Because HTML 5 is awesome. |
| 21:48 | <Philip`> | Presumably about:sgml-compat (or whatever it was going to be) would have to be defined as valid somewhere too |
| 21:48 | <Lachy> | ok. If it's not too difficult, I could write something up this weekend if I get motivated |
| 21:49 | <gsnedders> | Just use the XML format to write an RFC from scratch |
| 21:49 | <gsnedders> | It works fine. |
| 21:49 | <Lachy> | Philip`, I think only about:blank would need to be defined in the I-D. about:sgml-compat is just like any other. Any vendor or spec can mint their own about URIs |
| 21:50 | <Philip`> | Lachy: Would that need some IANA registry? |
| 21:51 | <annevk> | Hixie, yeah I see, not too difficult, although the XML format looks rather simple too |
| 21:51 | gsnedders | stretches his legs |
| 21:51 | <Lachy> | I don't know how to handle that. Since vendors should be able to freely mint their own about URIs as needed, but then it might be reasonable to require specs to register them |
| 21:51 | <gsnedders> | annevk: It is. |
| 21:52 | <Lachy> | though if we go down the registry route, I can see people complaining about the potential for clashes |
| 21:52 | <annevk> | can't we just say that about:<anything> is ok and that it's up to the "application layer" to say what happens and then HTML5 says what to do for about:blank |
| 21:53 | <annevk> | in the navigation section |
| 21:53 | <Hixie> | i'd rather not |
| 21:53 | <Hixie> | that's not really saying anything |
| 21:53 | <Hixie> | and we might end up with conflicting results |
| 21:54 | <Lachy> | is there an IANA registry of URI schemes somewhere? |
| 21:54 | <Lachy> | found it http://www.iana.org/assignments/uri-schemes.html |
| 22:03 | Philip` | wonders if anything uses the 'kern' tables in OpenType |
| 22:12 | <Dashiva> | Isn't kerning a pretty big deal in making fonts look pretty? |
| 22:13 | <Philip`> | Yes, but it seems everyone uses the GPOS table for that, rather than the kern table |
| 22:14 | <Philip`> | Libertine has a 0.75MB kern table, but if I delete the table then I can't see any difference for any text in any browser |
| 22:15 | <Dashiva> | Maybe it's used by font editing software to generate/update gpos :) |
| 22:18 | <Philip`> | kern is much more limited than GPOS, since it can only do pairs rather than arbitrary sequences, so that probably wouldn't be a terribly sane thing to do :-) |
| 22:18 | <Dashiva> | I'm out of ideas then |
| 22:19 | <Philip`> | The documentation just says "Windows v3.1 does not make use of the 'kern' data other than to expose it to applications through the GetFontData() API.Format 2" |
| 22:25 | <Lachy> | using urn:w3c:sgml-compat like Larry suggests, instead of about:sgml-compat could work. |
| 22:25 | <Lachy> | it's only mildly longer and almost as memorable |
| 22:28 | <annevk> | then use data:,sgml-compat |
| 22:30 | <Hixie> | sounds more formal |
| 22:30 | <Hixie> | i prefer about:sgml-compat personally |
| 22:30 | <Hixie> | data:,sgml-compat looks messy enough though :-) |
| 22:31 | <Hixie> | though that would confuse actual DTD-based processors somewhat :-) |
| 22:31 | <Lachy> | yeah, data URIs are messy. it's likely people will forget the comma |
| 22:31 | <annevk> | Hixie, they'd be in violation of HTTP! |
| 22:31 | <Hixie> | HTTP? |
| 22:32 | <Lachy> | I like about:sgml-compat too. But can't think of any good reason to object to the urn: |
| 22:33 | <Hixie> | what's the objection to about: ? |
| 22:33 | <Hixie> | it's shorter and simpler and looks less magical. |
| 22:33 | <annevk> | Hixie, trying to interpret a text/plain resource as DTD |
| 22:33 | <Lachy> | personally, I don't have an objection to about: |
| 22:34 | <Hixie> | annevk: that's not HTTP, that's MIME, but yeah, good point. |
| 22:36 | <Lachy> | annevk, http://www.w3.org/TR/html4/strict.dtd is served as text/plain |
| 22:41 | <annevk> | oooh, news at eleven, w3.org has bugs :p |
| 22:41 | <Dashiva> | Do what I publish, not what I do |
| 22:42 | <Lachy> | annevk, what bugs? |
| 22:42 | annevk | blinks |
| 22:42 | <Lachy> | my point was that serving a DTD as text/plain is ok |
| 22:43 | <Lachy> | application/xml-dtd is for XML DTDs only, not SGML DTDs, or fake DTDs |
| 22:43 | <annevk> | and that means they should use text/plain? |
| 22:44 | <Lachy> | what else should they use? |
| 22:44 | <annevk> | something unregistered? |
| 22:44 | <Lachy> | application/sgml maybe, which is registered |
| 22:45 | <annevk> | if content sniffing is bad, you can't put resources online as text/plain and expected them to be handled as anything but text |
| 22:46 | <Lachy> | meh |
| 22:50 | <annevk> | now personally I'm not sure content sniffing is bad and frankly the amount of effort it requires to get font/ttf and font/otf seems to be high so we'll probably be stuck with content sniffing for fonts |
| 22:55 | <Lachy> | there's no way we're going to be able to avoid content sniffing for fonts. |
| 22:56 | <Lachy> | well, it's not really content sniffing. It's more like treating anything retrieved from @font-face src will be treated as a font |
| 22:56 | <gsnedders> | Lachy: The sensible thing is to send no content-type |
| 22:57 | <Lachy> | gsnedders, the problem is that Apache sends a content type by default. Usually text/plain for unknown types |
| 22:57 | <gsnedders> | Lachy: Then application/octet-stream |
| 22:58 | <Lachy> | yeah, that's what I used when I made those @font-face TC's the other day |
| 23:21 | <Lachy> | gsnedders, does that mean you're not going to rewrite the introduction for my draft? |
| 23:22 | <gsnedders> | Lachy: Unless you really really really want me to, yes |
| 23:24 | <Lachy> | ok |
