| 00:00 | <Hixie> | i did a google search for [deprecated metadata vocabularies] |
| 00:00 | <Hixie> | and one of the top hits was a link to http://metadataregistry.org/conceptprop/show/id/13721.html |
| 00:01 | <Hixie> | which as far as i can tell is the registration page for a property in a vocabulary that describes vocabularies, to label a part of a vocabulary as deprecated. |
| 00:01 | <Hixie> | what the fuck. are you doing. |
| 00:10 | <jgraham> | Hixie: Can you register the registration vocabulary as deprecated in the registration vocabulary? |
| 00:20 | <Hixie> | jgraham: i don't see why not... |
| 00:27 | <TabAtkins> | Hixie: It's only valid *before* you actually do it. Once you write it, it becomes invalid (due to you using a deprecated vocab). |
| 00:32 | <annevk> | value objects <3 |
| 00:37 | <annevk> | Domenic_: what's the value from a snapshot? |
| 00:37 | annevk | is not a fan of people copying his work and creating mass confusion |
| 00:38 | <Domenic_> | annevk: patent protection presumably |
| 00:39 | <annevk> | A snapshot doesn't give you that |
| 00:41 | <Domenic_> | If taken to rec it does right? |
| 00:43 | <gsnedders> | Only REC gives you that. |
| 00:43 | <gsnedders> | The timeline in which you must withhold any patent grants is based upon the first WD with anything the patent pertains to. |
| 01:22 | <TabAtkins> | annevk: Yus. |
| 01:22 | <TabAtkins> | annevk: (re: value objects) |
| 01:23 | <TabAtkins> | I *really* want to write "el.css.width = 5px;" and have it magically work. |
| 01:23 | <TabAtkins> | (120px).pt == 90pt ! |
| 01:34 | <heycam> | TabAtkins, will value types get SVG out of the predicament of SVGAnimatedLengths etc. without having to boil the ocean like in my proposal? |
| 01:39 | <TabAtkins> | heycam: I don't think so. Value types are immutable objects that you can overload operators for. |
| 01:39 | <TabAtkins> | And define suffixes for, so "let x = 5px;" is equivalent to "let x = CSS.px(5);". |
| 01:39 | <heycam> | ok |
| 01:40 | <heycam> | how much sense would it be, if we do boil the ocean, make SVGRectElement.x be a CSS length value type rather than a DOMString then? |
| 01:40 | <TabAtkins> | It'll make for a *really convenient* CSSOM API, as you can just do "el.css.width = 5px;" |
| 01:40 | <TabAtkins> | It would make a lot of sense. |
| 01:40 | <TabAtkins> | But I'm expecting value types to take 1-2 years. |
| 01:40 | <TabAtkins> | So, shrug. |
| 01:41 | <heycam> | hmm |
| 01:41 | <TabAtkins> | annevk: If you make a URLValue class, I can use that directly in CSS rather than inventing a CSSURLValue object. |
| 02:45 | <annevk> | TabAtkins: yeah I guess that would make sense |
| 02:45 | <annevk> | TabAtkins: although I also think that having it as a string is okay or is there some ambiguity with that? |
| 02:59 | <Hixie> | 1-2 years! wow, that'd actually be pretty quick for something like that |
| 03:10 | <TabAtkins> | annevk-cloud: Nah, I need to distinguish them in the data model from actual strings. |
| 03:11 | <TabAtkins> | Hixie: Yeah, but the push for 64-bit ints has had years to mature, and this is riding on the back of that demand. |
| 03:13 | <Hixie> | cool |
| 08:41 | <Ms2ger> | I'm pretty sure the date on http://christianheilmann.com/2014/01/23/endangered-species-of-the-web-the-image/ is wrong |
| 08:41 | <Ms2ger> | It sounds more like a post from 1994 |
| 09:01 | <macfreek> | Good day! I like an account for wiki.whatwg.org, and as noted on top of the pages, I have to ask here for an account. |
| 09:02 | <macfreek> | My preferred username would be macfreek |
| 09:03 | <macfreek> | I like to exploit my neurotic behaviour and fix the markup error on the page MetaExtensions (there is a missing row break before the entry on revisit-after) |
| 09:05 | <macfreek> | Hmm, no-one here? |
| 09:05 | <macfreek> | I'll drop a mail at admin⊙wwo |
| 09:14 | <jgraham> | dglazkov: Thanks! |
| 09:32 | <MikeSmith> | IRC protip you may want to try hanging out for longer than 5 minutes if you're asking for help and nobody responds |
| 11:18 | <hsivonen> | Hmm. Does @theFakeW3C write for @w3cmemes, too? |
| 13:43 | <jgraham> | mathiasbynens: Nice work! |
| 13:45 | <zcorpan> | indeed :-) |
| 18:33 | <gsnedders> | So, the tokenizer is provably live-lock free. Which is good to know. But probably could've been sufficently shown with a lot less work. :) |
| 18:49 | <TabAtkins> | Oh good lord, i18n issue spam on CSS Text. |
| 18:52 | <annevk> | Domenic_: what do you think about https://www.w3.org/Bugs/Public/show_bug.cgi?id=24367 ? |
| 18:52 | <annevk> | Domenic_: always copy or keep a list of objects |
| 18:54 | <Domenic_> | annevk: what does "with an existing "url object" association" mean? why would they need to be associated? |
| 18:54 | <Domenic_> | oh i see why |
| 18:54 | <Domenic_> | so that changing them changes the URL |
| 18:54 | <annevk> | Domenic_: yeah, syncing |
| 18:54 | <annevk> | sync syncing, even |
| 18:57 | <Domenic_> | annevk: blah, no great answers I feel |
| 18:57 | <Domenic_> | maybe most pure would be url.searchParams.resetFrom(otherSearchParams) |
| 18:57 | <Domenic_> | and not allowing url.searchParams = ... |
| 18:58 | <Domenic_> | creating a copy has the usual `url.searchParams = x; url.searchParams !== x` problem |
| 18:59 | <annevk> | well at least url.searchParams === url.searchParams |
| 18:59 | <annevk> | I thought that was the invariant |
| 19:00 | <annevk> | what's the downside with always assigning? |
| 19:00 | <annevk> | and keeping a list of URL objects? |
| 19:00 | <Domenic_> | hmm |
| 19:02 | <Domenic_> | That actually works |
| 19:02 | <Domenic_> | I see no downside |
| 19:02 | <Domenic_> | I like it better than copying |
| 19:03 | <Domenic_> | as for invariants, well, you try to get as many of them as you can :) |
| 19:03 | <Domenic_> | as if it were just a normal data property not a getter/setter |
| 19:03 | <Domenic_> | x.y === x.y is the most obvious though, agreed. |
| 19:14 | <annevk> | I'm getting a 503 on GMail |
| 19:15 | <annevk> | I don't think I ever had Gmail down |
| 19:15 | <annevk> | Now it's a 500 |
| 19:21 | <annevk> | http://www.google.com/appsstatus#hl=en&v=issue&ts=1390636799000&sid=1&iid=8201b96db60ea3ccc26c53a4487c5492 |
| 19:21 | <MikeSmith> | annevk: https://twitter.com/oscargodson/status/426795522083401728 |
| 19:22 | <annevk> | Well, it does happen less than missile strikes |
| 19:23 | <annevk> | Not sure that makes it more newsworthy though |
| 21:05 | <Hixie> | MikeSmith: see e-mail to whatwg list from me just now |
| 21:19 | <annevk> | Domenic_: with streams, there's some kind of indication of new data being available right? |
| 21:19 | <annevk> | Domenic_: so you could implement progress events based on the content-length header plus listeners on the stream |
| 21:20 | <Domenic_> | annevk: yeah, you could definitely implement progress events on top of streams |
| 21:20 | <Domenic_> | as you read data, emit a progress event |
| 21:22 | <annevk> | Domenic_: so the idea is to push streams through TC39? |
| 21:22 | <Domenic_> | annevk: probably proceed the same way as with promises: proceed in web space, but try for TC39 space if there's interest and they are fast enough. Might be time for one of those supplemental spec things like i18n used |
| 21:23 | <Domenic_> | but, write the algorithms in TC39 spec style to smooth the process, and e.g. don't use DOMException |
| 21:24 | <annevk> | seems fair |
| 21:26 | <Hixie> | btw the e-mail i mentioned above to MikeSmith is actually something everyone might be interested in |
| 21:26 | <Hixie> | it's a discussion about what to do with registries, starting with <meta name>'s |
| 21:26 | <TabAtkins> | It was definitely an interesting email! |
| 21:26 | <Domenic_> | Hixie: it is a pretty interesting email. Sounds like a lot of work though! |
| 21:26 | <Hixie> | what would be a lot of work? |
| 21:27 | <Hixie> | actually writing the server side would be pretty easy, i imagine. the validator side would be more work. |
| 21:27 | <Hixie> | dunno how much work |
| 21:35 | <annevk> | It's on my todo list |
| 21:36 | <Domenic_> | I feel like the validator needs some advertisement on the main whatwg.org page |
| 21:37 | <Domenic_> | And maybe some rebranding. (X)HTML5 Living Validator is kind of weird |
| 21:38 | <annevk> | hsivonen: ^^ |
| 21:39 | <Hixie> | Domenic_: i can update the http://www.whatwg.org/ page -- what should i remove to replace with the validator link? |
| 21:39 | <Hixie> | (there's several things on that page commented out, i rotate the boxes around every year or so) |
| 21:41 | <Domenic_> | Forums seems worth killing... |
| 21:41 | <Ms2ger> | Heh, forums |
| 21:47 | <Hixie> | ok reload, i moved stuff around for 2014 |
| 21:48 | Ms2ger | only ever ends up there on accident |
| 21:48 | <Hixie> | yeah i dunno how many people load that page per se |
| 21:48 | <Hixie> | probably like 2 a month |
| 21:48 | <Hixie> | there's already a link to validators at the top of the HTML spec itself though |
| 21:49 | <annevk> | mailing list is close to 10 years http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2004-April/000000.html |
| 21:50 | <Domenic_> | Hixie: Nice new layout! FAQ gets the prominence it deserves... |
| 21:52 | <Hixie> | foolip: map, the usage of <source media> is stupendously low. i haven't seen numbers that low, like, ever. |
| 21:52 | <Hixie> | foolip: even features that don't _exist_ get more usage than that.! |
| 21:53 | <Hixie> | s/map/man/ |
| 21:53 | <Hixie> | Domenic_: yeah, like i said, i rotate things around :-) |
| 22:24 | <annevk> | https://pbs.twimg.com/media/BexMPGSCUAAeJzi.png:large |
| 22:34 | <annevk> | Domenic_: https://www.w3.org/Bugs/Public/show_bug.cgi?id=24367 got some more noise |
| 22:35 | <annevk> | Domenic_: I guess keep references to multiple owners is still the best |
| 22:35 | <annevk> | I did not mean noise as opposed to signal btw, just that more comments were added |
| 22:48 | <Domenic_> | annevk: commented and thus now subscribed. |
| 22:51 | Hixie | remembers how he hates bash scripting |
| 22:52 | <TabAtkins> | Hixie: So... don't? |
| 22:52 | <Hixie> | yeah |
| 22:53 | <Hixie> | it always starts as a trivial script that just runs three programs or whatever |
| 22:53 | <Domenic_> | i'd suggest shelljs but IIRC Hixie isn't a huge JS person either :P |
| 22:53 | <Hixie> | and then i add some piping... and then months later i need an if statement... next thing i know i'm trying to work out obscure [ syntax errors deep inside nested ``s |
| 22:54 | <TabAtkins> | The moment you advance beyond piping, switch to Python. |
| 22:54 | <TabAtkins> | Or Perl, or whatever you like. |
| 22:58 | <Hixie> | in principle i agree, but when you have a long sequence of system calls and just want to add one conditional, it's not an obvious choice. |
| 22:58 | <TabAtkins> | Yeah, so you have to set a Schelling Fence. |
| 22:58 | <TabAtkins> | Any single addition to the bash script doesn't seem like it's bad enough to be worth recoding. |
| 22:58 | <TabAtkins> | And then before you know it you're in hell. |
| 22:59 | <TabAtkins> | So just set an arbitrary reasonable point beforehand, and stick to it. |
| 23:01 | <Hixie> | i think i'm past that point |
| 23:01 | <Hixie> | looooong past that point |
| 23:02 | <TabAtkins> | So rewrite. ^_^ |
| 23:02 | <Hixie> | that point occured when i was a kid, learning that there were these tools that would do what i told them |
| 23:02 | <Hixie> | the bash vs not bash thing is just one point on the slope :-P |
| 23:05 | <Hixie> | anyway. i think a better way of approaching this particular problem isn't a schelling fence |
| 23:05 | <Hixie> | it's a deadline for how long i can afford to spend on fixing the problem in bash before i switch to another language |
| 23:06 | <Hixie> | so if i think i could rewrite it in, say, perl, in 1 hour, then maybe i would allow 30 minutes of bash debugging and then just switch over |
| 23:06 | <Hixie> | that way i don't waste an hour fixing something that took 1 minute |
| 23:12 | <TabAtkins> | *This* time. And next time you have a problem that's horrible to debug in bash, you'll make the same estimate, and probably come to the same conclusion. |
| 23:12 | <TabAtkins> | With the end result that you keep spending too much time painfully debugging bash, rather than eating the cost of a rewrite into a better language once. |
| 23:12 | <TabAtkins> | Thus, the Schelling Fence. |
| 23:21 | <Hixie> | TabAtkins: nah. that's only true if the code changes from being mainly running other scripts, to being mainly logic. |
| 23:21 | <Hixie> | TabAtkins: while it's mainly running other scripts and piping stuff around, bash is better than other languages. |
| 23:29 | <jgraham> | Hixie: https://pypi.python.org/pypi/plumbum |
| 23:29 | <jgraham> | Perhaps |
| 23:29 | <Hixie> | relying on non-default-install runtimes is a non-starter for me |
| 23:30 | <jgraham> | …OK |
| 23:30 | <Hixie> | i end up spending way too much time reinstalling libraries and so on |
| 23:30 | <Hixie> | i've set up a schelling fence around it. :-) |
| 23:31 | <jgraham> | Generally package management isn't a huge issue |
| 23:31 | <jgraham> | I find |
| 23:32 | <jgraham> | I mean this isn't 1998 where you have to compile stuff from source all the time |
| 23:32 | <jgraham> | Unless you are on OSX I guess |
| 23:32 | <jgraham> | Which is just awful |
| 23:38 | <Hixie> | jgraham: it becomes way more of a pain when you're trying to not do anything as root (including installing packages) |
| 23:40 | <jgraham> | Hixie: I just virtualenv all python stuff |
| 23:40 | <Hixie> | yeah if i did more python stuff i should probably get that figured out |
| 23:40 | Hixie | isn't a huge python fan so hasn't bothered to learn it much |
| 23:41 | <Hixie> | sorry if it sounds like i'm full of stop energy today :-) |
| 23:41 | <jgraham> | If I had a bunch of bash-like stuff in python (I don't) I would just have a venv in my home directory that would be activated in my .bash_login |
| 23:41 | <Hixie> | a lot of this is automated stuff, but yeah |
| 23:41 | <Hixie> | (cron jobs, cgi scripts, etc) |
| 23:41 | <jgraham> | Then external packages would just be "pip install foo" and no need for root |
| 23:57 | <SimonSapin> | Hixie: is the multipage version generated asynchronously? I get the "This specification has been updated." message, even after reloading |
| 23:59 | <SimonSapin> | jgraham: you could add venv/bin to $PATH without "activating" it |
