| 00:06 | <boblet> | anyone know the status of HTML5+RDFa validation? coming soon? being worked on? not yet? |
| 00:07 | <hober> | IIRC hsivonen isn't working on that |
| 00:08 | <hober> | but fortunately v.nu is open source, so perhaps RDFa advocates will build such a thing |
| 00:10 | <boblet> | hober: thanks for that. I noticed that w3.org has XHTML+RDFa so wondered if it might be coming… |
| 02:47 | <Hixie> | annevk: yt? |
| 04:31 | <Hixie> | . o |
| 04:31 | <Hixie> | ('>) |
| 04:31 | <Hixie> | LL |
| 04:32 | Hixie | sacrifices the chicken to summon annevk or abarth |
| 04:33 | <othermaciej> | poor chicken :-( |
| 04:34 | <Hixie> | abarth asked for XMLDocument to be specced (with .async and .load()) |
| 04:34 | <Hixie> | i'm hoping i can convince anne to spec it in XHR |
| 04:35 | <othermaciej> | hmm |
| 04:35 | <Hixie> | but in their absence, i'm having to spec it myself |
| 04:35 | <Hixie> | curse the time zone difference! |
| 04:35 | <othermaciej> | I think the primary use is via createDocument() not XHR |
| 04:36 | <Hixie> | oh it has nothing to do with XHR except that it'd reuse a lot of his infrastructure |
| 04:36 | <Hixie> | it has nothing to do with HTML either |
| 04:36 | <othermaciej> | I also once again idly wonder if it is "compatible enough" to omit the sync version |
| 04:36 | <Hixie> | but there we go |
| 04:36 | <Hixie> | yeah i saw your mail about that |
| 04:36 | <Hixie> | i have no data either way |
| 04:36 | <Hixie> | currently webkit browsers don't do it at all |
| 04:36 | <othermaciej> | because synchronous I/O is the devil |
| 04:36 | <Hixie> | so in some sense, we can omit it altogether... |
| 04:37 | <othermaciej> | abarth provided fairly compelling evidence that it is a nontrivial compatibility loss |
| 04:37 | <Hixie> | yes |
| 04:37 | <othermaciej> | we also found compelling evidence that you can't just add the load() method to all documents |
| 04:37 | <Hixie> | yes |
| 04:37 | <othermaciej> | by hunting around in old mozilla bugs |
| 04:37 | <othermaciej> | no data specific to the sync version afaik |
| 04:45 | <GPHemsley> | Is it intended behavior for it to be impossible to style the weight of the text in <option>? |
| 04:46 | <Hixie> | i don't understand how document.load() interacts with DOM manipulation of the document in question |
| 04:47 | <Hixie> | it seems to be racy |
| 04:49 | <Hixie> | in particular, document.load() empties the DOM of the document object, but then if you append a node immediately, what should happen? |
| 04:50 | <Hixie> | anyone got IE? what does IE do with something like http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3E%0D%0A%20%20%20var%20xmlDoc%20%3D%20document.implementation.createDocument(%22%22%2C%20%22%22%2C%20null)%3B%0D%0A%20%20%20w('A%3A%20'%20%2B%20xmlDoc.documentElement)%3B%0D%0A%20%20%20xmlDoc.appendChild(xmlDoc.createElement('test'))%3B%0D%0A%20%20%20w('B%3A%20'%20%2B%20xmlDoc.documentElement)%3B%0D%0A%20%20%20xmlDoc.onload%2 |
| 04:50 | <Hixie> | make that http://software.hixie.ch/utilities/js/live-dom-viewer/saved/575 |
| 04:59 | <Hixie> | blimey, xhr is non-trivial these days |
| 05:43 | <Hixie> | othermaciej: looks like Firefox's sync document.load() isn't so synchronous |
| 05:43 | <Hixie> | othermaciej: it blocks script but nothing else |
| 05:43 | <Hixie> | http://www.hixie.ch/tests/adhoc/dom/level0/document/load/002.html |
| 05:45 | <othermaciej> | Hixie: it looks like it is running a nested event loop |
| 05:45 | <Hixie> | yeah |
| 05:45 | <othermaciej> | that's unfortunate |
| 05:46 | <othermaciej> | I guess it's nice that it doesn't block the UI, but it surely creates all sorts of re-entrancy challenges |
| 05:47 | <Hixie> | yeah |
| 05:47 | <othermaciej> | it also violates JS run-to-completion semantics |
| 05:47 | <othermaciej> | Hixie: suggested experiment: make that w('LOAD START') into an alert and see what happens if you close the window before the load completes |
| 05:47 | <cardona507> | anyone going to the GTUG campout at Google HQ in a couple of weeks? |
| 05:48 | <cardona507> | the focus is HTML5 :D |
| 05:49 | <Hixie> | othermaciej: http://www.hixie.ch/tests/adhoc/dom/level0/document/load/003.html |
| 05:49 | <Hixie> | othermaciej: it immediately fires the alert, with no window |
| 05:50 | <Hixie> | othermaciej: interesting way of working around onunload limitations |
| 05:50 | <Hixie> | othermaciej: (in other words, it closes the window, and as part of that terminates all network activity, which then resumes the script) |
| 05:51 | <othermaciej> | Hixie: I get no LOAD START alert at all, but I might be testing with an older Firefox |
| 05:52 | <Hixie> | othermaciej: what version are you testing? |
| 05:52 | <Hixie> | i have Mozilla/5.0 (Macintosh; Intel Mac OS X 10.5; rv:2.0b3pre) Gecko/20100724 Minefield/4.0b3pre |
| 05:53 | <othermaciej> | Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.6) Gecko/20100625 Firefox/3.6.6 |
| 05:53 | <Hixie> | ah, a release |
| 05:53 | <Hixie> | weird that they would have changed this since then |
| 05:53 | <Hixie> | but oh well |
| 05:59 | <othermaciej> | probably not an intentional change |
| 06:00 | <othermaciej> | I bet the script runs to completion both ways, and the difference is just whether a windowless alert fails silently or does something |
| 06:20 | <annevk> | hmm, responseXML is meant to cache |
| 06:21 | <annevk> | though it can also change if you invoke overrideMimeType |
| 06:39 | <hsivonen> | Hixie: shouldn't we go for removing .load() from the platform instead of speccing it? |
| 06:39 | <hsivonen> | Hixie: if you use .load() in Firefox 4, it warns to console that the feature is going away |
| 06:42 | <mcarter> | Hixie, haha, good catch wrt that email... I just resent |
| 06:44 | <annevk> | Hixie, aren't the fetch flags <i> rather than <var>? |
| 06:48 | <annevk> | Hixie, the simple event needs to be queued in the async case I think |
| 06:49 | <annevk> | Hixie, and in the sync case what happens if you invoke load() from the event handler? |
| 06:49 | <annevk> | Hixie, might actually also be interesting for the async case |
| 06:51 | <annevk> | Hixie, also, http://twitter.com/WHATWG/statuses/19633786262 links one revision to far? |
| 07:05 | <hsivonen> | whoa? is this real? https://blogs.apache.org/foundation/entry/the_apache_software_foundation_receives |
| 07:06 | <hsivonen> | I didn't know that non-profits can be sold to for-profits |
| 07:07 | <hsivonen> | ah. April 1 getting replayed by a planet |
| 07:07 | <MikeSmith> | hey, it's mcarter |
| 07:07 | <MikeSmith> | long time no see |
| 07:08 | <mcarter> | yeah, hoping to be more of a help with WebSocket stuff |
| 07:08 | <MikeSmith> | way cool |
| 07:08 | <MikeSmith> | great to see back around |
| 07:09 | <mcarter> | I wet my hands a bit with http://orbited.org/orbited2 recently, and it got me pretty excited about actually having websocket native |
| 07:12 | <MikeSmith> | mcarter: I think there was a Nostradamus prophecy that said you would re-appear when WebSocket was in a time of great need |
| 07:12 | <mcarter> | hahaha |
| 07:13 | <mcarter> | The only real great need is for someone to hit reply "what are you talking about? no." to every second hybi email over at the ietf |
| 07:16 | <MikeSmith> | heh |
| 07:20 | <Hixie> | hsivonen: i'd love to... remove it from firefox and let me know how that goes |
| 07:21 | <Hixie> | annevk: oops (re <var>/<i>) |
| 07:21 | <Hixie> | annevk: the load event is queued, the whole set up substeps is queued in fact |
| 07:22 | <Hixie> | annevk: didn't try reentrancy, good question |
| 07:24 | <annevk> | oh sorry, missed the queue a task bit before |
| 07:24 | <boblet> | so with lots of phrasing content elements also being flow content, is it still “good practice” to wrap a flow/phrasing element e.g. <small> in a formatBlock candidate e.g. <p>? I know <footer><small> works, but I’d assumed a <p> wrapper was added to DOM… |
| 07:25 | <annevk> | I now wonder whether I should further intertwine XMLHttpRequest by using the same definition of "XML MIME type" and "Content-Type metadata" |
| 07:25 | <Hixie> | annevk: (the tweet being wrong is because i screwed up the checkin) |
| 07:26 | <annevk> | boblet, yeah, implied paragraph semantics is just a convenience |
| 07:27 | <Hixie> | http://www.hixie.ch/tests/adhoc/dom/level0/document/load/005.html - i don't understand what firefox is doing |
| 07:27 | <annevk> | Hixie, and there's a ~ after step 1 I now notice |
| 07:27 | <boblet> | annevk: cool, thanks. “best practice” is hard |
| 07:28 | <Hixie> | annevk: yeah, i fixed that |
| 07:29 | <annevk> | boblet, initially some wanted to outlaw <body> <img> <p> test </p> </body> and such (like HTML4) but iirc hsivonen convinced Hixie that allowing it would be easier and still ok (with the implied paragraphs concept) |
| 07:35 | <boblet> | annevk: initially that looked fine to me, but img are display:inline by default huh. implied <p> is definitely author-friendly |
| 07:36 | <annevk> | boblet, well yeah, we no longer call them inline or block level either you might have noticed |
| 07:36 | <annevk> | boblet, notice that implied paragraph does not mean a <p> in the DOM |
| 07:37 | <annevk> | boblet, it just means that semantically a paragraph is there, for user agents that care about paragraph semantics |
| 07:37 | <boblet> | annevk: yes to both :) although initially I thought flow was an equivalent to block-level |
| 07:40 | <annevk> | for the new element groupings it probably helps to not think about HTML4 at all, though I guess that will be rather hard for at least another couple of years |
| 07:40 | <boblet> | annevk: and if you’re trying to explain them to ppl who only know HTML 4 |
| 07:41 | <boblet> | btw John Allsopp had the brainwave of using [role] instead of the JS shiv for IE7-8. still need it for IE6 of course |
| 07:43 | <annevk> | together with the XHTML2 WG he's probably the only role="" fanatic out there who doesn't consider it an attribute for assistive technology |
| 07:43 | <annevk> | iirc he even got an ALA article on that published |
| 07:44 | <boblet> | annevk: http://twitter.com/johnallsopp/status/19617327472 ;-) |
| 07:44 | <boblet> | he’s reached “in a perfect world…” equilibrium |
| 07:46 | <annevk> | ah, well, welcome to the future |
| 07:47 | <boblet> | but not the future of the perfect world |
| 07:59 | <annevk> | boblet, perfect worlds don't have angle brackets |
| 08:00 | <boblet> | annevk: guess we’ll have to settle for this world |
| 08:02 | <othermaciej> | hola |
| 08:03 | <othermaciej> | hsivonen: I think it would be better to get rid of it, but abarth discovered pretty serious Web compat consequences to not having it |
| 08:08 | <erlehmann> | othermaciej, that sentence alone does adequately describe a whole lot of spec quirks ;) |
| 08:09 | <othermaciej> | hsivonen: I would love it if Mozilla would actually remove it, or at least the synchronous version - but i don't know if a console warning carried much weight |
| 08:10 | <annevk> | we would like that too |
| 08:13 | <annevk> | would be nice if there were some people writing papers on how to do format extensibility on the Web, why versioning is an anti-pattern on the Web, etc. |
| 08:13 | <annevk> | for future spec writers |
| 08:16 | <annevk> | maybe when my beard turns gray |
| 08:17 | <Hixie> | annevk: yeah i am surprised as to how often such basic concepts are found to not be understood by people at the w3c or ietf |
| 08:17 | <Hixie> | annevk: yet they keep telling us that the reason to go to the w3c or ietf is that we'll get more expert review |
| 08:18 | <annevk> | that reminds me of something funny I read in the HyperText CG this morning |
| 08:19 | <annevk> | first couple of sentences from http://www.w3.org/2010/07/16-hcg-minutes.html#item01 |
| 08:31 | <hsivonen> | annevk: that minute item is sad |
| 08:34 | <micheil> | Hixie: I always do that on the list. >_> |
| 08:34 | <micheil> | I'm use to the google group's mailing lists where the reply-to is the mailing list email address. |
| 08:34 | <Hixie> | don't feel bad, mcarter did the exact same thing like 15 minutes before you :-P |
| 08:35 | <micheil> | heh heh |
| 08:35 | <micheil> | I've done that a few times now; oh well. |
| 08:36 | <micheil> | Hixie: is there anything important I should mention on that podcast about the work being done on the websocket protocol? |
| 08:38 | <Hixie> | no casualties so far, but the battles are still raging so it's too early to count our chickens? |
| 08:40 | <hsivonen> | count chickens? |
| 08:40 | <Hixie> | that was a very mixed metaphor |
| 08:41 | <Hixie> | the chicken part was an allusion to the proverb "don't count your chickens until the eggs have hatched" |
| 08:43 | <hsivonen> | ah |
| 08:52 | <mcarter> | is xhr.sendAsBinary a whatwg thing, or an api that mozilla made up? |
| 08:54 | <annevk> | the latter |
| 08:54 | <annevk> | XHR is a W3C thing atm |
| 08:59 | <mcarter> | is there a draft of the proposed binary apis (for webgl, xhr, and potentially websocket) that I saw hixie mention in his recent hybi email? |
| 08:59 | <annevk> | anyone have a link handy to all the polls? |
| 08:59 | <annevk> | ah |
| 08:59 | <annevk> | http://www.w3.org/2002/09/wbs/40318/ |
| 08:59 | <Hixie> | http://www.w3.org/2002/09/wbs/40318/issue-101-objection-poll/ http://www.w3.org/2002/09/wbs/40318/issues-4-84-objection-poll/ |
| 08:59 | <MikeSmith> | mcarter: https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/doc/spec/TypedArray-spec.html is one |
| 09:00 | <MikeSmith> | mcarter: if that's what you meant |
| 09:00 | <mcarter> | MikeSmith, thanks, I think it may be what I'm talking about |
| 09:01 | <MikeSmith> | mcarter: there are some links at the commonjs wiki as well |
| 09:01 | MikeSmith | tries to find the page |
| 09:01 | <MikeSmith> | http://wiki.commonjs.org/wiki/Binary |
| 09:01 | <othermaciej> | the Khronos folks met with ECMA TC-39 at some point but I don't know what progress has been made |
| 09:01 | <mcarter> | I wonder if it would be beneficial to create a binary WebSocket api that could accept a string OR one of these ArrayBuffers. For the string case it would just do what xhr.sendAsBinary does now (truncate the higher order byte of each utf-16 character) |
| 09:02 | <MikeSmith> | othermaciej: it's good news that they met at least |
| 09:02 | <MikeSmith> | would be nice if they made some minutes available |
| 09:04 | <annevk> | hah, public minutes of either alone is already hard afaik |
| 09:05 | <Hixie> | ok bed time |
| 09:05 | <Hixie> | nn |
| 09:07 | <mcarter> | goodnight Hixie |
| 09:20 | <annevk> | answered the surveys... |
| 09:20 | <annevk> | feels like a waste of time, especially the 101 |
| 09:25 | <annevk> | still no news on licensing :/ |
| 10:46 | <jgraham> | Philip` or someone: where are the instructions for committing to the HTML testsuites repository? |
| 10:52 | <jgraham> | (I found it) |
| 10:55 | <annevk> | use https and your W3C username/password |
| 10:55 | <annevk> | I think there's not much else to it |
| 10:55 | <annevk> | at least not when I played with the WebApps repository the other day |
| 10:57 | <jgraham> | Yeah, it was straightforward in the end |
| 11:07 | <hsivonen> | Reading the remote XUL removal bug, I wonder how people decide to develop single-browser apps (be it IE, Firefox or Safari) |
| 11:07 | <hsivonen> | well, I guess there probably are Opera-only apps, too |
| 11:07 | <hsivonen> | widgets probably |
| 11:08 | <annevk> | I read some of that bug, not sure what to think of it |
| 11:09 | <Philip`> | I'd guess it's the same way they decide to develop single-OS apps |
| 11:11 | <Philip`> | The platform provides features they want to use, enough of their potential users are running that platform, the platform is relatively stable so their app isn't going to break soon, etc |
| 11:14 | <hsivonen> | why is it so hard to get non-Google webmail that doesn't suck? |
| 11:14 | <hsivonen> | Zimbra is regressing |
| 11:14 | <hsivonen> | now it no longer fills in "Foo Bar wrote:" when replying to a message. |
| 11:19 | <ksemeks> | how do i in chrome enlarge the databaseStorage size? i m trying with javascript, but the size is always the same |
| 11:37 | <othermaciej> | hsivonen: we have learned about a lot of content that is either WebKit-specific or WebKit/Gecko-specific due to the recent HTML5 malformed close tag parsing issue |
| 11:38 | <othermaciej> | content that depends on WebKit bugs/quirks/proprietary features includes: Dashboard widgets, apps that use WebKit for their display engine (like chat programs), mobile-targeted versions of web sites, Apple enthusiast centric web sites, and Apple-internal servers |
| 11:42 | <othermaciej> | (I think it's specifically the changed handling of <abc</abc> - we will probably have to put a quirk in the HTML5 parser to restore behavior for WebKit-based native apps at minimum) |
| 11:42 | <Philip`> | (And Apple web standards marketing demos) |
| 11:43 | <hsivonen> | othermaciej: also, <canvas> centric stuff (Bespin) and Gecko-specific intranets (Mozilla's own!) |
| 11:43 | <Workshiva> | Google is luckily not a monoculture there :) |
| 11:43 | <hsivonen> | that is, Bespin relied on </foo</bar> |
| 11:44 | <hsivonen> | and at least two Mozilla intranet pages have relied on recovery when </title> is missing |
| 11:44 | <othermaciej> | <abc</abc> is unusual because HTML5 goes with the IE/Opera behavior instead of the WebKit/Gecko behavior |
| 11:44 | <Philip`> | http://philip.html5.org/data/gt-in-tag.txt |
| 11:44 | <othermaciej> | maybe that's not all that unusual |
| 11:44 | Philip` | can't remember whether he sent that link to abarth or not |
| 11:44 | <othermaciej> | but it would be nice to know if there is real breakage from the WebKit/Gecko behavior |
| 11:44 | <annevk> | there's breakage either way |
| 11:44 | <Philip`> | Um, I should have called that lt-in-tag.txt |
| 11:46 | <hsivonen> | (AFAIK, Mozilla's intranet doesn't intend to be Gecko-specific, but stuff that doesn't work in IE but works in Firefox and Mobile Safari happens.) |
| 11:47 | <othermaciej> | I guess we'd never find out the real impact of breakage the other way until IE implements the HTML5 parsing algorithm |
| 11:47 | <hsivonen> | othermaciej: I believe Opera has experience of that breakage |
| 12:28 | <annevk> | so http://www.w3.org/Bugs/Public/show_bug.cgi?id=9745 means the WebSocket handshake is going to change? |
| 13:02 | <jgraham> | annevk: Maybe for port 443 only? |
| 13:02 | <jgraham> | or rather wss only? |
| 13:03 | <annevk> | I suspect it will be something like that |
| 15:12 | <AryehGregor> | Finally, <https://bugs.webkit.org/show_bug.cgi?id=40747>; is fixed. |
| 15:13 | <hsivonen> | aargh. https://bugs.webkit.org/show_bug.cgi?id=40747#c3 |
| 15:13 | <AryehGregor> | I hope it gets widely deployed to Safari before anyone has a decent implementation that anyone would actually want to use. |
| 15:13 | <hsivonen> | can someone, please, stop the Chrome team from adding new doctype sniffing? |
| 15:14 | <AryehGregor> | They did, see: https://bugs.webkit.org/show_bug.cgi?id=40520 |
| 15:14 | <hsivonen> | AryehGregor: ok. whew. |
| 15:15 | <AryehGregor> | Initially tkent implemented it for all modes, then saw some bugs and disabled it in quirks mode, then people yelled at him for making the modes different, and people also yelled at him for doing validation at all without UI, so now he's just disabled it across the board. |
| 15:15 | <AryehGregor> | And when UI is added, presumably it will be readded for all modes. |
| 15:15 | <AryehGregor> | Although by the way, I think that by "<!doctype html>" he means "non-quirks". |
| 15:18 | <jgraham> | Argh. Don't do that to me |
| 15:18 | <jgraham> | People adding mode-specific things makes me want to cry blood |
| 15:19 | <AryehGregor> | People adding things that are completely broken and unusable and thereby burdening web developers with having to do browser sniffing makes me want to cry sulfuric acid. |
| 15:19 | <AryehGregor> | (which is much more painful) |
| 15:25 | <jgraham> | Who's to say my blood tears were't caused by initial sulfuric acid tears corroding the veins in my eyes? |
| 15:29 | <AryehGregor> | But at least in that case, the tears are restricted to your eyes, whereas mine will cause terrible burns across my whole face, and also my hands when I frantically try to wipe it away, and my mouth when it leaks in as I scream. |
| 15:29 | <AryehGregor> | Also, wouldn't the acid cauterize the veins? |
| 15:30 | <Philip`> | The terrible pain sounds like a bonus, as it will distract you from the HTML issues that initiated the tears |
| 15:31 | <annevk> | what is a WG Secretary? |
| 15:35 | <annevk> | http://tools.ietf.org/html/rfc2418#section-6.2 -- minute taker |
| 15:42 | <MikeSmith> | Lachy: can you remind what the status is on the Selectors API spec? |
| 15:56 | <hsivonen> | reminder for everyone who has commented on ARIA: the deadline for accepting/rejecting responses in the current round is "by" tomorrow |
| 15:56 | <hsivonen> | (I'm going through my responses now...) |
| 16:50 | <TabAtkins> | cardona507: I think I'll be at the campout as a counselor. |
| 16:51 | <cardona507> | sweet - I am going to be there also - |
| 16:51 | <cardona507> | how did your talk go the other day? |
| 16:54 | <cardona507> | TabAtkins: what are your thoughts on node.js? |
| 16:58 | <TabAtkins> | My talk went fine. The two responses we got back, one very negative and one very positive, were both very accurate. ^_^ |
| 16:58 | <TabAtkins> | I think node.js is interesting. I haven't touched it, but I've read up on it a little bit. |
| 17:04 | <cardona507> | any slides from the talk? I am doing front end work on a rails project and the team is interested in nodejs for an upcoming project so I am just now becoming aware of nodejs. It seems pretty cool. And there is definitely a buzz around it. |
| 17:05 | <TabAtkins> | Not from me - I basically just reused my demos from my last talk. The other dudes had some slides, though I dunno where they are. |
| 17:25 | <miketaylr> | cardona507: if you're anywhere near NYC theres a node.js hackday/workshop this saturday |
| 17:25 | <cardona507> | miketaylr: I am in Cali - thanks for the heads up though :) |
| 17:27 | <TabAtkins> | Ooh, on that note, we've got an internal tech talk about node.js tomorrow! Yay! |
| 17:27 | <miketaylr> | cardona507: its only like a 7 hour flight, cmon! ;) |
| 17:28 | <cardona507> | TabAtkins: was that your subtle way of inviting us? ;) |
| 17:28 | <cardona507> | miketaylr: if only it were that easy... |
| 17:30 | <miketaylr> | yeah, it's only a 30 minute train ride for me and it's not that easy |
| 17:30 | <TabAtkins> | cardona507: I dunno if outsiders can attend. I'll ask, if you want. |
| 17:30 | <jgraham> | Dear "I'm too lazy to read the spec right now and will instead go home"web: is there a simple way to get the DOMContentLoaded event on the document on an iframe when you set the src of the iframe (so you don't have the right document object at the point where you can set an event listener) |
| 17:31 | <cardona507> | TabAtkins: please ask - that would be great |
| 17:44 | <cardona507> | miketaylr: will any of the talks be streamed this saturday? |
| 17:44 | <miketaylr> | cardona507: i doubt it. i think its more of a hands-on hack day |
| 17:44 | <cardona507> | cool |
| 17:48 | <gsnedders> | jgraham: You were talking about ORMs coping cleanly with cases where some conditions were optional, got any code of one doing so? |
| 17:48 | <gsnedders> | To ask again… |
| 17:48 | gsnedders | weeps at jgraham's silence |
| 18:02 | <jgraham> | gsnedders: Either I am ignoring you or you missed "and will now go home" |
| 18:02 | <jgraham> | gsnedders: No examples, but it seems like it shouldn't be too hard |
| 18:03 | jgraham | reads some documentation |
| 18:04 | <jgraham> | gsnedders: Is what you want different from using sql alchemy and doing |
| 18:05 | <jgraham> | f = filter(foo.bar = "baz") |
| 18:05 | <jgraham> | if condition1: |
| 18:05 | <jgraham> | f = f.filter(foo.foobar = "something") |
| 18:05 | <jgraham> | and so on? |
| 18:05 | <gsnedders> | no |
| 18:06 | <gsnedders> | But I thought you meant there was a way cleaner than that :) |
| 18:06 | <jgraham> | Oh |
| 18:06 | <jgraham> | Well that seems clean relative to gluing together bits of SQL by hand depending on the conditions :) |
| 18:07 | <jgraham> | (and you can probably make it look nicer if you have some good way of mapping conditions to filters) |
| 18:07 | <gsnedders> | Well, in the cases where you just create an array and concat them with OR it's not massively cleaner |
| 18:09 | <jgraham> | I tend to think that anything not involving gluing together strings is a win |
| 18:16 | <AryehGregor> | TabAtkins, in ISSUE-4/ISSUE-84, aren't you supposed to quote things rather than just linking to them? |
| 18:18 | <TabAtkins> | AryehGregor: I linked, but then summarized the takeaway point from the link in my reply. |
| 18:41 | <hsivonen> | TabAtkins: I think your objecting could use some restatement of the argument for readers who don' already agree with robin's conclusion |
| 18:41 | <hsivonen> | *objection |
| 18:59 | AryehGregor | notices this while reading the HTTP 1.0 spec: "Product tokens [like the User-Agent header] should be short and to the point -- use of them for advertizing or other non-essential information is explicitly forbidden. Although any token character may appear in a product-version, this token should only be used for a version identifier (i.e., successive versions of the same product should only differ in the product-version portion of the pro |
| 18:59 | <AryehGregor> | Ha. Ha. Ha. |
| 19:00 | <TabAtkins> | I blame it all on early hackers. Fuckers, the lot of 'em. |
| 19:01 | <hsivonen> | yay for reality-based specs |
| 19:05 | <AryehGregor> | Proxy Error |
| 19:05 | <AryehGregor> | The proxy server received an invalid response from an upstream server. |
| 19:05 | <AryehGregor> | The proxy server could not handle the request POST /2002/09/wbs/40318/issues-4-84-objection-poll/. |
| 19:05 | <AryehGregor> | Reason: Error reading from remote server |
| 19:05 | <AryehGregor> | I get that when trying to submit my objection at: http://www.w3.org/2002/09/wbs/40318/issues-4-84-objection-poll/ |
| 19:05 | AryehGregor | pokes MikeSmith? |
| 19:06 | <AryehGregor> | (If that had caused me to lose my giant post that I spent like an hour writing, I'd have killed someone.) |
| 19:06 | <Philip`> | Maybe you wrote too much |
| 19:06 | <Philip`> | I think hsivonen had that problem |
| 19:07 | <AryehGregor> | (Browsers need more bulletproof auto-form-saving. It's reliable enough these days that I rely on it, but just unreliable enough that it doesn't work enough times to be kind of frustrating.) |
| 19:07 | Philip` | suggests killing the people who haven't added a 'recover recently submitted form values' feature to browsers |
| 19:07 | <AryehGregor> | $ wc ~/issue4-objection |
| 19:07 | <AryehGregor> | 32 929 5804 /home/aryeh/issue4-objection |
| 19:07 | <AryehGregor> | Is that too much? |
| 19:08 | <Philip`> | 13:01 < MikeSmith> if a comment is 5629 characters or longer, submitting a response will fail |
| 19:10 | <Philip`> | (You could Twitterise your response so it fits in the limit, or post to www-archive and link to it) |
| 19:10 | <AryehGregor> | Seriously? 5629 is the cutoff? |
| 19:10 | <AryehGregor> | Why 5629? |
| 19:11 | Philip` | shrugs |
| 19:11 | <AryehGregor> | Wikipedia used to have a problem where some browsers wouldn't handle text boxes properly if they had 32 KB or more of text, but that's at least a vaguely comprehensible number. |
| 19:12 | AryehGregor | is used to Twitterizing git commit summaries so the first line is 50 characters or less, so that should be no big deal to do |
| 19:14 | <AryehGregor> | Well, seems like it worked. |
| 19:14 | <AryehGregor> | I wasn't too far over, so no big deal. |
| 19:18 | <hsivonen> | AryehGregor: I had the same problem. MikeSmith instructed me to post to www-archive and point from the survey to www-archive |
| 19:18 | <AryehGregor> | hsivonen, I was only a couple hundred bytes over, so I just shortened it a bit. |
| 19:32 | <jgraham> | AryehGregor: Maybe the real limit is 8192 or something, but a whole load of extra bytes are added somewhere? |
| 19:32 | <AryehGregor> | Would make sense. |
| 19:32 | <jgraham> | 8192 is still kinda radom though |
| 19:32 | <jgraham> | +n |
| 19:33 | <AryehGregor> | I'm guessing the system wasn't really stress-tested so much prior to the HTMLWG "everyone is a member" philosophy. |
| 19:33 | <jgraham> | With this poll it's more like "everyone is Alexander Pope" |
| 19:34 | <othermaciej> | AryehGregor: if you can't make your response short enough, I you can always mail it to www-archive and post a link |
| 19:34 | <AryehGregor> | othermaciej, it's okay, I just shortened it a bit. |
| 19:35 | jgraham | decides that Pope was not quite who he had in mind |
| 19:35 | <jgraham> | Although if anyone does write their response in verse form, I shall be impressed |
| 19:36 | <jgraham> | Possibly the chairs less so :) |
| 20:36 | <Hixie> | Philip`: your advice on http://www.w3.org/Bugs/Public/show_bug.cgi?id=9574 would be welcome |
| 20:47 | <jgraham> | Yay for floating point differences |
| 20:50 | <Philip`> | Hixie: I have no idea how to define anything that would lead to more interoperability than leaving it undefined |
| 20:51 | <gsnedders> | jgraham: No, you're just working in the wrong base. |
| 20:54 | <Philip`> | If it's defined like "if |endAngle-startAngle| < 0.001 then assume they are the same point" then you're just shifting the problematic boundary from 0 to 0.001 (e.g. what if |end-start| = 0.00100001) |
| 20:55 | <TabAtkins> | Define the boundary as a base-2 fraction. |
| 20:56 | <Philip`> | What if I set startAngle = 0, endAngle = 1000*Math.PI? |
| 20:57 | <Philip`> | Are we going to define the precise 64-bit floating-point value of Math.PI and mod operation and all the rounding modes and everything? |
| 20:57 | <TabAtkins> | That's what we'd need for interop, presumably. |
| 20:57 | <TabAtkins> | "dyadic fraction" is the word I was looking for. |
| 20:57 | <Philip`> | I'd be surprised if everyone was willing to implement things that precisely |
| 20:58 | <TabAtkins> | Me too. |
| 20:58 | <Philip`> | particularly since changing rounding modes would hurt performance |
| 20:58 | <AryehGregor> | Does x87 have different rounding from the new fancy better x86 floating-point stuff? |
| 20:59 | <AryehGregor> | (we can always just say ARM is non-conforming) |
| 20:59 | <Philip`> | I don't know how to avoid problems without either defining everything in way too much precision (so most implementations won't bother to follow it), or defining functions so they don't have discontinuities in the output as a function of their inputs |
| 20:59 | <TabAtkins> | We could define the cutoff to be obnoxiously large, so rounding is easier. "Anything smaller than 1/256 is rounded to 0". |
| 21:00 | <AryehGregor> | Then you have to do explicit rounding all the time in the implementation. |
| 21:00 | <AryehGregor> | Will anyone be willing to do that? |
| 21:00 | <Philip`> | (It'd probably be possible to define arc that way but I think the current definition was discussed at length before being settled on, and it's not much fun to do that all over again) |
| 21:02 | <Philip`> | AryehGregor: If I remember correcty, x87 uses the rounding mode specified by the FPU control word, which sometimes gets fiddled with by parts of the application or by drivers etc when they want performance |
| 21:03 | <AryehGregor> | Awesome. |
| 21:03 | <Philip`> | so you need to be quite careful to keep control of that |
| 21:05 | <Philip`> | On amd64 you just use SSE for all floating point |
| 21:05 | <Philip`> | which has its own rounding mode flag |
| 21:05 | <gsnedders> | AryehGregor: x87 also uses 100-bit fp, whereas SSE2 uses 32/64-bit fp |
| 21:05 | <TabAtkins> | 100b? That sounds weird. |
| 21:05 | <Philip`> | Also, compilers will happily rearrange all your FP operations and keep things in 80-bit registers and mess up your calculations |
| 21:05 | <gsnedders> | So if you do multiple instructions on a single piece of data, rounding differences can appear |
| 21:05 | <TabAtkins> | Are they using 28b for something else? |
| 21:05 | <AryehGregor> | I take it that any remotely recent x86-32 CPUs has SSE too, but the browsers don't want to use it because then they'll break on everyone's old Pentium II's? |
| 21:05 | <Philip`> | unless you pass lots of restrictive compiler flags and hope the optimiser respects them sufficiently |
| 21:05 | <Philip`> | gsnedders: 80, not 100 |
| 21:05 | <AryehGregor> | TabAtkins, 80-bit floating point is common too. |
| 21:05 | <gsnedders> | AryehGregor: SSE2 is P4 at lesat |
| 21:05 | <gsnedders> | Philip`: Gah, close enough. |
| 21:05 | <gsnedders> | != 64 is the point |
| 21:05 | AryehGregor | guesses "too" was inappropriate, then |
| 21:05 | <jgraham> | gsnedders: wrong base? |
| 21:06 | <gsnedders> | AryehGregor: (and SSE2 introduced double percision fp) |
| 21:06 | <gsnedders> | jgraham: well, it's not my fault that 0.1 + 0.7 != 1.0 |
| 21:06 | <TabAtkins> | 80b still sounds weird. It's even further from it's smallest larger power of 2. |
| 21:06 | <Philip`> | http://mxr.mozilla.org/mozilla-central/source/js/src/jsnum.cpp#664 is setting x87 to use 64-bit instead of 80-bit, I think |
| 21:08 | <Philip`> | TabAtkins: "Anything smaller than 1/256" - how does that help? It just means your boundary is now 1/256, not 0/256 |
| 21:08 | <gsnedders> | http://code.google.com/p/v8/issues/detail?id=436 |
| 21:08 | <TabAtkins> | Philip`: It helps because 1/256 is a precise value in all extant and expected future fp implementations. |
| 21:09 | <Philip`> | TabAtkins: So is 0 |
| 21:11 | <TabAtkins> | Oh, hrm, you're right. I'm dumb. |
| 21:11 | <AryehGregor> | TabAtkins, not everything has to be in powers of two, you know. |
| 21:11 | <TabAtkins> | 1/256+e will cause the exact same problems. |
| 21:11 | <TabAtkins> | AryehGregor: Heresy! |
| 21:11 | <Philip`> | https://bugzilla.mozilla.org/show_bug.cgi?id=531915 - those functions are even worse (differences in the ~6th digit) |
| 21:13 | <AryehGregor> | Math.tan(-1e300) = -4.802497308567303 (Linux), -4.987183803371025 (Windows) |
| 21:14 | <AryehGregor> | That's awful. |
| 21:14 | <AryehGregor> | Although not too awful when you think about what you're asking. |
| 21:14 | <gsnedders> | ooo, fireworks! |
| 21:24 | <jgraham> | Also, fireworks! |
| 21:24 | <jgraham> | Although I didn't se them and they have stopped now |
| 21:27 | <gsnedders> | THE INTERNET IS BREAKING! |
| 21:27 | <gsnedders> | Or at least that's what I feel like every time there's a netsplit. |
| 21:27 | <gratz|home> | why did you break it? |
| 21:27 | <TabAtkins> | Philip`: If you define it as "everything less than 1/2^n rounds to 0", the problem of being an epsilon over 1/2^n disappears, but being an epsilon *under* 1/2^n remains, right? |
| 21:29 | <annevk> | after more than a year: http://www.w3.org/TR/2010/WD-cors-20100727/ |
| 21:30 | <Philip`> | TabAtkins: Depends what problem you mean |
| 21:31 | <Philip`> | The actual problem isn't about rounding to zero, it's about whether you're above or below 2pi |
| 21:31 | <TabAtkins> | In that 1/2^n - epsilon may be equal to or less than 1/2^n, depending on accuracy. |
| 21:32 | <TabAtkins> | Oh, right, that's a separate and interesting thing. |
| 21:32 | <TabAtkins> | We should express it in piradians rather than radians. |
| 21:32 | <TabAtkins> | Then you can just draw from 0 to 2. |
| 21:32 | <TabAtkins> | Or turns! |
| 21:33 | <Philip`> | Oh, actually, doesn't the spec handle this already? |
| 21:33 | <Philip`> | I forgot to read one of its paragraphs |
| 21:40 | Philip` | keeps confusing himself |
| 21:44 | Philip` | gives up since his brain is incapable of doing geometry |
| 21:48 | <hsivonen> | hmm. something is wrong with taking snapshots by year http://www.w3.org/News/2010.html#entry-8864 |
| 21:49 | <hsivonen> | if the 2007 snapshot goes to LC in 2010 |
| 21:50 | <TabAtkins> | Agreed. |
| 21:52 | <annevk> | I wondered if that was intentional |
| 21:52 | <annevk> | not that it matters much |
| 21:52 | <jgraham> | Intentional to make it look like a silly waste of time? |
| 21:55 | <Philip`> | Putting dates in names seems a terrible idea, because you'll always miss them |
| 21:56 | <Philip`> | even if you take C++0x's approach of giving yourself a 10-year interval |
| 21:56 | <TabAtkins> | We could have just renamed it to 2010 snapshot! |
| 21:56 | <TabAtkins> | Publishing a 2010 snapshot in 2011 still sounds reasonable. |
| 21:57 | <Philip`> | Call it the 2012 snapshot and say you're ahead of schedule |
| 21:57 | <annevk> | Snapshot Beijing is cooler |
| 21:57 | <TabAtkins> | Problem with that is that the name is completely opaque. Which can be a good or bad thing, I guess. |
| 21:57 | <annevk> | 2007 is also opaque |
| 21:57 | <Philip`> | Names should be opaque identifiers |
| 21:57 | <Philip`> | and should make it clear that they are |
| 21:58 | <annevk> | and not completely; Beijing was where this was decided |
| 21:58 | <TabAtkins> | By "opaque" I meant "meaningless to anyone not intimately familiar with the CSSWG's operation and the locations in which their FtFs take place". |
| 21:58 | <annevk> | in 2007 |
| 21:58 | <hsivonen> | x in C++0x is now hex |
| 21:59 | <annevk> | would be fun to have a series of specs named after beers |
| 21:59 | <TabAtkins> | I could get behind that. |
| 22:00 | <hsivonen> | does anyone want to place bets on when all the platforms that browser engines need to be ported to have robust C++0x compilers? |
| 22:00 | <Philip`> | hsivonen: Doesn't really make much sense to count in hex the number of years after 2000 which is only significant because of its decimal representation |
| 22:00 | <Philip`> | hsivonen: Are "the platforms" basically GCC + MSVC? |
| 22:01 | <TabAtkins> | Philip`: Nah, 0x is the standard *prefix* for hex numbers, thus the year is completely unspecified now. |
| 22:01 | <hsivonen> | Philip`: various version of GCC, GCC without standard lib (Android), Sun Studio |
| 22:01 | <hsivonen> | maybe clang in near future |
| 22:02 | <Philip`> | Is GCC without standard lib considered a robust C++98 compiler? |
| 22:02 | <hsivonen> | it's sad that my attitude towards C++0x is so similar to the attitude of IE6-supporting authors towards CSS3 |
| 22:03 | <hsivonen> | Philip`: probably not, but browser engines definitely need to compile for Android |
| 22:03 | <Philip`> | Lots of the C++0x libraries are already in Boost, so you can use them by just choosing a different namesace |
| 22:03 | <Philip`> | s//p/ |
| 22:04 | <hsivonen> | Boost has an odd license, IIRC |
| 22:07 | <Philip`> | http://www.boost.org/users/license.html doesn't look too odd to me |
| 22:08 | <hsivonen> | Philip`: looks like a one-off license to me |
| 22:09 | <Philip`> | It's an n-off licence where n is the number of Boost libraries |
| 22:10 | <franksalim> | lgtm: http://www.opensource.org/licenses/bsl1.0.html, http://www.gnu.org/licenses/license-list.html |
| 22:10 | <hsivonen> | well, my point is that to use it, you need to ping licensing gatekeepers and everyone downstream needs to ping theirs |
| 22:10 | <hsivonen> | which sucks big time compared to using a well-known license that has already been approved for use everywhere |
| 22:11 | <Philip`> | Are any already approved for use everywhere? |
| 22:11 | <Philip`> | (Maybe MIT or BSD?) |
| 22:11 | <hsivonen> | MIT and 3-clause BSD are doing pretty well |
| 22:12 | <TabAtkins> | And GPL, within an ecosystem that's okay with it. |
| 22:12 | <jgraham> | I would be nice if everyone used MIT | GPL |
| 22:12 | <jgraham> | | LGPL I guess |
| 22:12 | <franksalim> | BSD > MIT |
| 22:12 | <jgraham> | Why? |
| 22:13 | <AndresBotero> | Public domain > BSD |
| 22:13 | <franksalim> | the endorsement clause |
| 22:13 | <jgraham> | The endorsement clause doesn't make much sense to me |
| 22:13 | <jgraham> | I don't see why you need it |
| 22:13 | <Hixie> | me either |
| 22:14 | <Hixie> | i think MIT is better than BSD-3-clause specifically because of that clause |
| 22:14 | <hsivonen> | non-endorsement is the default anyway |
| 22:15 | <jgraham> | I would have thought that if someone was misrepresenting your name there would be plenty of other avenues to go down |
| 22:15 | annevk | is with AndresBotero as far as specs/tests go |
| 22:16 | <TabAtkins> | Indeed. |
| 22:16 | <franksalim> | I can see how bsl, although it is very similar to mit/bsd, could cause you a minor headache if you need legal approval for uncommon licenses. on the other hand, if you just want to make sure you can get your packages into debian, it's fine |
| 22:16 | TabAtkins | duals his stuff in public domain and cc0. |
| 22:17 | franksalim | isn't writing a web browser, though... |
| 22:17 | <jgraham> | I was under the impression that public domain scared some people |
| 22:17 | <AndresBotero> | It is really about "tastes" and future plans. |
| 22:18 | <TabAtkins> | Problem is that public domain doesn't actually exist in some countries, or means something different. |
| 22:18 | <AndresBotero> | Well, some people just want recognition, others just want the world to know about what they create. |
| 22:18 | <TabAtkins> | So I'm putting trust in CC to get CC0's licensing right. |
| 22:19 | <franksalim> | cc0 has an endorsement clause |
| 22:19 | <franksalim> | there must be some reason for that |
| 22:19 | <Hixie> | cc has done a lot of things wrong imho :-) |
| 22:19 | <gsnedders> | AndresBotero: Some people just want help with what they're working on for themselves if anyone else cares :) |
| 22:19 | <TabAtkins> | Hixie: Agreed! |
| 22:20 | <TabAtkins> | franksalim: you mean the "When citing the work, you should not imply endorsement by the author." sentence? |
| 22:20 | <TabAtkins> | (and associated legalese translation) |
| 22:21 | <franksalim> | yes, but i actually can't find it in the legalese version |
| 22:21 | <jgraham> | The CC0 pages only seem to show the legal text in a tiny box when you click "begin" |
| 22:22 | <jgraham> | You are then asked if you agree with all the stuff in the tiny box |
| 22:22 | <TabAtkins> | franksalim: You're right - I cant' find it in the legal code. |
| 22:22 | <jgraham> | That is insane |
| 22:22 | <TabAtkins> | http://creativecommons.org/publicdomain/zero/1.0/legalcode |
| 22:24 | <jgraham> | TabAtkins: So how would one find that? The only way I could do it was to load the iframe on the signup page as a tlbc |
| 22:24 | <TabAtkins> | jgraham: I dunno how you'd find it from the signup, but from the link they give you it's a link near the bottom: |
| 22:24 | <TabAtkins> | http://creativecommons.org/publicdomain/zero/1.0/ |
| 22:25 | <jgraham> | I didn't even find that page |
| 22:26 | <TabAtkins> | AFter you ask for a license, it should give you that link. |
| 22:30 | <jgraham> | Anyway, for software MIT seems simple and safe |
| 22:30 | <TabAtkins> | Yeah. |
| 22:33 | <Hixie> | why just software? |
| 22:36 | <jgraham> | Hixie: Because it is specifically a software license |
| 22:37 | <jgraham> | If you change it to not be about software, it is no longer the MIT license (although it may be effectively the same) |
| 22:38 | <jgraham> | The indemnification part doesn't make much sense for e.g. a novel though |
| 22:43 | <Hixie> | jgraham: the MIT license is about software and documentation |
| 22:43 | <Hixie> | specifically, "this software and associated documentation files" |
| 22:45 | <jgraham> | OK, but it implies that there is a non-zero amount of software involved |
| 22:45 | <jgraham> | I don't think you could use it unmodified to releease *just* documentation |
| 22:45 | <AndresBotero> | Yeah, not specifically about JUST software, or just documenttation. |
| 22:45 | <AndresBotero> | That |
| 22:52 | <Hixie> | jgraham: since the empty string is a valid program in many languages, i don't think that's a problem. There's software in between each letter of the documentation! :-) |
| 22:52 | <TabAtkins> | So, your book is just software written in cat? |
| 22:53 | <Hixie> | cat? |
| 22:53 | <TabAtkins> | The utility. |
| 22:53 | <Hixie> | that's not a programming language |
| 22:53 | <Hixie> | i use perl |
| 22:54 | <TabAtkins> | Pretty sure there's an esolang named that, which used cat as its interpreter. |
| 22:54 | <Hixie> | ah |
| 22:54 | <TabAtkins> | http://esoteric.voxelperfect.net/wiki/Cat_program |
| 22:54 | <Hixie> | well i meant "real" programming languages |
| 22:54 | <TabAtkins> | And http://esoteric.voxelperfect.net/wiki/Text |
| 22:56 | <AryehGregor> | TabAtkins, #!/usr/bin/cat is unlikely to work, unless you're on Solaris. |
| 22:56 | <daedb> | There's also http://en.wikipedia.org/wiki/Whitespace_(programming_language) of course :) |
| 22:58 | daedb | wonders what would happen if you fed the html spec into a whitespace interpreter... |
| 23:30 | <Hixie> | i can't work out if this guy is called "Lobotom Dysmon" (as per his sig) or "Thomas Bouchardon" (as per his From: line) |
| 23:30 | <TabAtkins> | I'm leaning toward the latter. |
| 23:30 | <othermaciej> | the latter sounds a little more like a real name |
| 23:31 | <Hixie> | only the former has a facebook page |
| 23:32 | <TabAtkins> | That's not a real account, though. |
| 23:32 | <Hixie> | i can't tell |
| 23:32 | <Hixie> | since i don't have one myself :-) |
| 23:33 | <TabAtkins> | It exists for the sole purpose of displaying that avatar. |
| 23:34 | <volkmar> | Hixie: Thomas Bouchardon sounds like a typical french name |
| 23:34 | <Hixie> | yes |
| 23:37 | <Philip`> | Hixie: It looks like you've got both names in the acknowledgements now |
| 23:37 | <Hixie> | d'oh |
| 23:37 | <Hixie> | oh well |
| 23:37 | <Philip`> | I suppose that's one way to be more certain of getting the right one, but I'm not sure whether it's intentional |
| 23:38 | <Hixie> | it'll do |
| 23:39 | <Philip`> | Also you have a Gavin before a Gareth |
| 23:41 | <gavin> | gavins are better than gareths |
| 23:42 | <Philip`> | Being later in the list is clearly better than being earlier |
| 23:44 | <AryehGregor> | People with Japanese names get better billing, because their names stand out. |
| 23:45 | <AryehGregor> | Maybe I should ask for my name there to be changed to "אריה גרגור (Aryeh Gregor)". |
| 23:45 | <Philip`> | You should send some feedback under a fake email address with a Japanese name that's actually a rude joke |
| 23:46 | <AryehGregor> | Alternatively, I should keep in mind to name all my children "Aaaaaaaaaaaaaah" so they get top billing in situations like this. |
| 23:46 | <AryehGregor> | (possibly with an exclamation point at the end) |
| 23:47 | TabAtkins | notes that he must name his children Aaaaaaaaaaaaaaah so as to beat Aryeh's children. |
| 23:47 | <Philip`> | There's a game called AaaaaAAaaaAAAaaAAAAaAAAAA!!! |
| 23:47 | <AryehGregor> | I'm in pretty good shape, all told, since it's sorted by first name rather than last. Too bad my parents didn't name me Aaron, then I'd be number three. Although if they did, I might go by Aharan, which would spoil it all. |
| 23:47 | <AryehGregor> | Philip`, I think I saw the review in PC Gamer. |
| 23:47 | Philip` | quite likes it |
| 23:48 | <Hixie> | i'm very happy to put in your name in its original form if you tell me what it is :-) |
| 23:48 | <AryehGregor> | TabAtkins, then I will simply prefix all my children's names with a null byte, since of course that sorts before everything. |
| 23:48 | <AryehGregor> | (I just relied on that for MediaWiki, in fact) |
| 23:48 | <Hixie> | AryehGregor: a null byte will turn into FFFD and end up at the end :-P |
| 23:48 | <TabAtkins> | "You perform stunts, weaving around the bustling City for points, making split-second decisions: do you snake around those girders to earn a dozen “kisses,” or glide along the side of that steel super-skyscraper for massive “hugs”?" |
| 23:48 | <AryehGregor> | (http://www.mediawiki.org/wiki/Special:Code/MediaWiki/69977) |
| 23:49 | <TabAtkins> | Prefix your kids' names with a BEEP |
| 23:49 | <AryehGregor> | Hixie, only if you use processing tools that aren't binary-safe! I just need to make sure all my e-mails are sent with MIME type application/octet-stream. |
| 23:49 | <AryehGregor> | Then if you mangle the null bytes, you're flagrantly violating the relevant standards. |
| 23:49 | <Philip`> | I never got many points from hugs - kisses seem much more important for points |
| 23:50 | <Philip`> | and points are important because then you win teeth |
| 23:51 | TabAtkins | will prefix his kid's names with U+-1. |
| 23:51 | <TabAtkins> | THE LOST CODEPOINT |
| 23:52 | <TabAtkins> | D'oh, I'm not in the acknowledgements. |
| 23:54 | <Hixie> | hah |
| 23:54 | <Hixie> | i noticed that the other day |
| 23:54 | <Hixie> | i figured i'd add you next time i came across an e-mail from you that caused me to change the spec, since that's my normal policy |
| 23:54 | <Hixie> | but i haven't come across one yet :-( :-P |
| 23:54 | TabAtkins | wonders if every email he's ever sent has been rejected. |
| 23:54 | <Hixie> | maybe! |
| 23:55 | <Hixie> | more likely i just missed adding your name at some point |
| 23:55 | <Hixie> | and then thought i'd added you so missed many more afterwards |
| 23:55 | <TabAtkins> | At the very least, I've influenced your decisions through the medium of IRC. |
| 23:56 | <TabAtkins> | Dude, 4 days ago you fixed a typo based on my input. |
| 23:56 | <TabAtkins> | Well, hmm. I suppose it was actually Alex Bishop who spotted it. |
| 23:58 | <TabAtkins> | Here we go. Your "Forms feedback" email from April 6th had you making a change based on a suggestion from me. |
| 23:59 | <Hixie> | lies! i, i, i thought of it first! |