| 00:40 | <rubys> | $ curl --head http://www.whatwg.org/ |
| 00:40 | <rubys> | curl: (7) couldn't connect to host |
| 00:41 | <Hixie> | yeah i'm rebooting it |
| 00:41 | <Hixie> | it should be up momentarily |
| 00:41 | <rubys> | cool, thx |
| 00:43 | <Hixie> | the server's been acting flaky for a couple of weeks now, but i can't work out exactly what's wrong which makes it hard to report |
| 00:44 | <rubys> | while I hadn't thought deeply about it until now, I guess I'm amazed that it is just one server. |
| 00:44 | <Hixie> | it's a VM that's hosting a dozen or so domains running on a shared server that itself is probably running dozens of VMs |
| 00:46 | <Hixie> | ok, i've escalated this to a proper outage request, it's not responding to my requests to reboot |
| 00:46 | <Hixie> | i expect some disk died or something |
| 00:47 | <Hixie> | ok, it's up again |
| 00:48 | <Hixie> | ish |
| 00:49 | <rubys> | lol. I noticed that some things seem sluggish... the first thing I noted was the simple serving of /images/logo took a several seconds. Subsequent requests were fine. |
| 00:51 | <rubys> | ... anyway, I assume that you are on top of it, and it will get resolved (either fixed where it is, or the site will move to a more reliable host, or whatever) |
| 01:32 | <nessy> | is the whatwg server hosted at Google? just curious.. |
| 01:39 | <Philip`> | nessy: Dreamhost |
| 01:40 | <nessy> | ah, ok - I expected Google to manage having servers up all the time, but I kinda do that from Dreamhost, too - anyway, good luck debugging the problems! |
| 06:54 | <Hixie> | is it normal for a php.cgi process to be long lived? |
| 06:55 | <Hixie> | php5.cgi, sorry |
| 06:58 | <annevk5> | i suppose it could be if it's doing heavy processing |
| 06:58 | <annevk5> | but I believe typically they're done rather fast |
| 06:59 | <Hixie> | that's what i thought |
| 06:59 | <Peter`> | lighttpd tends to spawn one of these for each process |
| 06:59 | <Hixie> | some of the blog php5.cgis and some of the wiki php5.cgis have been running for more than 15 minutes |
| 06:59 | <Peter`> | depends on the way the httpd communicates with them I guess, are they using resources? |
| 07:00 | <Hixie> | over 60 megs of RAM each |
| 07:03 | <annevk5> | maybe we should get some more caching for the blog? |
| 07:04 | <annevk5> | gruber makes fun of taking down wordpress blogs all the time too... so not sure if that'd help |
| 07:07 | <Hixie> | the problem isn't the blog |
| 07:07 | <Hixie> | there's something fishy with the server |
| 07:07 | <annevk5> | oh, maybe today the XHR test suite will start to work |
| 07:47 | <annevk5> | btw, about this registry stuff |
| 07:48 | <annevk5> | the XPointer people didn't use IANA either and have set up a registry that is pretty much what we want |
| 07:48 | <annevk5> | http://www.w3.org/2005/04/xpointer-schemes/ |
| 07:48 | <annevk5> | http://www.w3.org/2005/04/xpointer-policy.html |
| 07:48 | <annevk5> | "first come, first served", put directly in the registry, and then subject to further review |
| 07:49 | <annevk5> | URL to a detailed spec is optional |
| 07:49 | <annevk5> | it requires a W3C account, but everyone can get one of those |
| 07:50 | <Hixie> | i'd prefer a system that put less of a load on w3c staff and relied more on an open community, but otherwise yeah, that's pretty much what we need |
| 07:56 | <othermaciej> | I'm going to suggest to my fellow co-chairs that it's time for alternate / counter proposals, so if anyone wants to suggest a registry in that style, that would make a reasonable proposal |
| 07:56 | <othermaciej> | that does seem like a much more lightweight approach |
| 07:57 | <annevk5> | right, I'd still prefer a wiki over the thing above |
| 07:58 | <annevk5> | if someone prefers nice export et al they can write a wiki-to-JSON bridge |
| 07:59 | <othermaciej> | if anyone wants to propose using the wiki that is fine too |
| 07:59 | <othermaciej> | I don't care that much how this all works |
| 07:59 | <othermaciej> | but I think a wiki sucks for purposes of consumption the validator |
| 07:59 | <othermaciej> | and the validator is one of the most important consumers for this data |
| 08:00 | <othermaciej> | I would not expect a random wiki page to retain a sufficiently consistent format that you could soundly screen-scrape it to JSON |
| 08:04 | <annevk5> | Henri already wrote a wiki scraper for UI strings |
| 08:04 | <annevk5> | on purpose, since he could have hosted them in the repository or some such instead |
| 08:05 | <annevk5> | so I think that concern is rather theoretical |
| 08:06 | <annevk5> | a wiki has an auth system, it has lots of people looking at it, it deals with spam, etc. |
| 08:06 | <othermaciej> | it seems needlessly sloppy to me |
| 08:06 | <othermaciej> | there's really no benefit to the wiki page being a free-form editable piece of text |
| 08:06 | <othermaciej> | it doesn't really need the ability to be anything but a table with the appropriate N columns or whatever |
| 08:06 | <annevk5> | sure, but having to write new software to do everything a wiki already does is no benefit either |
| 08:07 | <othermaciej> | I could also imagine a setup where you use a wiki page for provisional registration, and something more structured for full "approved" registration |
| 08:07 | <othermaciej> | since the later is the only part that affects conformance |
| 08:07 | <othermaciej> | well, in theory anyway |
| 08:07 | <othermaciej> | IIRC the HTML5 spec does not define how the contents of the wiki affect conformance |
| 08:09 | <othermaciej> | hmm I guess it does say now |
| 08:10 | <othermaciej> | and it says both "proposed" and "ratified" count as conforming |
| 08:11 | <othermaciej> | strikes me as more generous than necessary (fail to validate, edit wiki, step 3 profit?) |
| 08:11 | <othermaciej> | but as I said, I don't care much about the details here |
| 08:12 | <annevk5> | proposed should prolly be a warning of some sorts |
| 09:09 | <Hixie> | othermaciej: there is one benefit to it being a free-form editable piece of text -- the code to do that already exists. |
| 09:09 | <Hixie> | and is deployed |
| 09:09 | <Hixie> | and has an active spam-fighting community |
| 09:09 | <Hixie> | (hober) |
| 09:09 | <othermaciej> | sure, and the HTML4.01 spec is already written and been published as REC |
| 09:10 | <othermaciej> | there's lots of things we could choose not to do because something exists |
| 09:10 | <othermaciej> | not saying this case is of nearly equal importance |
| 09:10 | <Hixie> | sure, i'm just disagreeing with "there's really no benefit to the wiki page being a free-form editable piece of text" |
| 09:10 | <Hixie> | i'm happy to have something better |
| 09:11 | <Hixie> | but unless someone steps up and writes the code for it, that's theoretical. |
| 09:11 | <annevk5> | and maintains it |
| 09:11 | <Hixie> | indeed |
| 09:12 | <Hixie> | (i think it would be better to have a wiki on the w3c's site than the whatwg site, btw) |
| 09:12 | <Hixie> | (wikipedia or wikibooks would be even better if they were up for it) |
| 09:13 | <annevk5> | maybe they can create wikistry |
| 09:13 | <Hixie> | hehe |
| 09:14 | <othermaciej> | that would probably deliver more of a promise of long-term continuity and be politically more congenial to some |
| 09:15 | <othermaciej> | still strikes me personally as too unstructured |
| 09:15 | <othermaciej> | but I have already talked way too much about a topic I profess to care little about |
| 09:15 | <Hixie> | heh |
| 09:21 | <jgraham> | Hixie: We can leave the innerHTML-on-foreign-content stuff to "v2" of course, but I would expect it to be in demand given that many js libraries rely heavily on innerHTML |
| 09:22 | <jgraham> | and will presumably be used in more situations with mixed namespace content once it is supported in text/html |
| 09:30 | <Hixie> | jgraham: yeah, agreed |
| 09:36 | <Hixie> | ok bed time |
| 09:36 | <Hixie> | nn |
| 09:37 | <annevk5> | nn |
| 09:37 | <jgraham> | gn |
| 09:55 | <MikeSmith> | gsnedders: I think I drink about a gallon of green tea before lunch |
| 09:55 | <MikeSmith> | and two gallons after |
| 09:56 | <gsnedders> | MikeSmith: I normally drink a lot of green tea, but today has been black |
| 09:56 | <MikeSmith> | I can't drink that much black tea |
| 09:57 | <MikeSmith> | it gives me ick mouth after a point |
| 09:57 | <MikeSmith> | and if I drank that much coffee, i'm pretty sure I'd have to be hospitalized |
| 09:57 | <gsnedders> | ewww, coffee |
| 09:58 | <MikeSmith> | gsnedders: btw, I get the impression you have lost interest, but I will soon be unleashing on you an enhancement patch for anolis |
| 09:58 | <MikeSmith> | and if nothing else it should provide you with some laughing-at-my-expense entertainment value |
| 09:59 | <gsnedders> | MikeSmith: I've said as much |
| 09:59 | <MikeSmith> | well, at this will olet you get a glimpse of my, um, approach to coding |
| 09:59 | <gsnedders> | MikeSmith: I'll probably dump the code I have for Anolis 2 somewhere public soon and just let anyone who wants xdoc support to get it to work |
| 09:59 | <MikeSmith> | OK, well, I'll have something you can add to the dump |
| 10:00 | <MikeSmith> | and possibly something you can contribute to the Daily WTF and get an award or something |
| 10:01 | gsnedders | hugs MikeSmith |
| 10:01 | <gsnedders> | Self-confidence! |
| 10:01 | <MikeSmith> | gsnedders: your code is nice and refined, like a jeweler working on a watch movement |
| 10:01 | <gsnedders> | MikeSmith: … |
| 10:01 | <MikeSmith> | mine is like a guy bashing a TV on the side to make it show a better picture |
| 10:01 | <gsnedders> | MikeSmith: Anolis was the first thing I ever wrote in Python. It's horrific. |
| 10:02 | <MikeSmith> | heh |
| 10:02 | <MikeSmith> | you ain't seen nothing yet |
| 10:02 | <gsnedders> | I've written cleaner Perl than the Python of Anolis. |
| 10:02 | <gsnedders> | Oh, I've seen plenty of bad code. |
| 10:02 | <MikeSmith> | I will say that this exercise of trying to add a new feature to anolis has given me a whole new appreciation for the etree API |
| 10:03 | <MikeSmith> | which is not saying much, because my appreciation for that API was previously zero |
| 10:04 | <MikeSmith> | and I would not have ever tried to use it unless it seemed like the quickest and bash-the-TV-iest way to get done what I wanted to get done |
| 10:04 | <MikeSmith> | but now that I've messed around with it a bit, I have to say it has its own appeal |
| 10:04 | <MikeSmith> | kind of like Bizarro Superman |
| 10:04 | <gsnedders> | It has some fairly horrible flaws, though |
| 10:05 | <MikeSmith> | true that |
| 10:05 | <jgraham> | The etree api as implemented in lxml is pretty good |
| 10:05 | <jgraham> | Could do with a .textContent |
| 10:05 | <MikeSmith> | it doesn't have one of its own? |
| 10:06 | <MikeSmith> | .text_content() ? |
| 10:06 | <MikeSmith> | or is that part of lxml? |
| 10:06 | <MikeSmith> | ah yeah |
| 10:06 | <jgraham> | Umm, maybe it does now |
| 10:06 | <gsnedders> | You can't do anything better than etree.to_string(node, type="unicode") |
| 10:06 | <gsnedders> | oh, lxml.html does |
| 10:06 | <gsnedders> | Odd |
| 10:06 | <jgraham> | "".join(node.xpath(".//text()")) |
| 10:06 | <gsnedders> | jgraham: Slow, comparitively |
| 10:07 | <gsnedders> | (Anolis used to do that) |
| 10:07 | <MikeSmith> | i recall this thing: |
| 10:08 | <MikeSmith> | etree.tostring(foo, method='text') |
| 10:08 | <gsnedders> | return etree.tostring(Element, encoding=unicode, method='text', with_fail=False) is what Anolis does now |
| 10:08 | <MikeSmith> | I think I used that etree.tostring(foo, method='text') for about 3 seconds, before I found some right-er way of doing it |
| 10:09 | <gsnedders> | anolislib.utils.textContent(foo) |
| 10:09 | <gsnedders> | :P |
| 10:09 | <MikeSmith> | anyway, it's a nice respite for the pain of working with native W3C DOM directly |
| 10:10 | <MikeSmith> | s/respite for/respite from/ |
| 10:11 | <MikeSmith> | the entire W3C DOM should be enshrined in Daily WTF in a place of high honor |
| 10:11 | <annevk5> | hmm, adding more features to XMLHttpRequest Level 2 while change requests are ongoing for Level 1 |
| 10:11 | <annevk5> | this is gonna be fun |
| 10:11 | <annevk5> | maybe I should wait |
| 10:13 | <annevk5> | hmm, I think that leaves CSS values |
| 10:13 | <annevk5> | I don't really like working on that |
| 10:13 | <MikeSmith> | W3C DOM Daily WTF award: The guy who couldn't conceive that path/selector node selection is way better the constant redundant tree walking all the freaking time |
| 10:13 | <MikeSmith> | what is CSS values? |
| 10:13 | <MikeSmith> | a political document? |
| 10:13 | <MikeSmith> | that should be a lot of fun to work on |
| 10:14 | <annevk5> | its like ele.style.values.width.px++ |
| 10:14 | <MikeSmith> | ah |
| 10:14 | <annevk5> | it's even |
| 10:15 | <MikeSmith> | even? |
| 10:15 | <annevk5> | but I guess I should at least take a stab at it, working on XHR2 can wait a little longer, and I'm not sure if I want to start on Web DOM Core just yet |
| 10:15 | <annevk5> | it's rather than its |
| 10:16 | <annevk5> | correcting a typo |
| 10:16 | <MikeSmith> | ah |
| 10:16 | <MikeSmith> | Web Notifications fun will be ramping up pretty soon |
| 10:16 | <annevk5> | I think Web DOM Core might be my next challenge in cleaning up legacy stuff |
| 10:16 | <MikeSmith> | yeah, do that |
| 10:16 | <annevk5> | ah yeah, that will be my challenge at adding cruft to the platform that creates future legacy problems ;p |
| 10:17 | <MikeSmith> | heh |
| 10:18 | <MikeSmith> | "and people thought pop-up windows *within* their browser were annoying..." |
| 10:18 | <MikeSmith> | that's what people will be saying next year |
| 10:18 | <annevk5> | hopefully we do it right |
| 10:18 | <annevk5> | harhar |
| 10:19 | <MikeSmith> | at least there's no additional telcons |
| 10:19 | <MikeSmith> | hopefully |
| 10:20 | <annevk5> | as long as I remain chair that seems unlikely |
| 10:21 | <MikeSmith> | I like your chairing already |
| 10:21 | <Peter`> | lol |
| 10:21 | <MikeSmith> | keep it up |
| 11:27 | <annevk5> | complicated: http://wiki.csswg.org/spec/cssom |
| 11:28 | <annevk5> | hmm |
| 11:29 | <annevk5> | now i wonder what license the IANA registry is under |
| 11:29 | <annevk5> | i guess that issue was never solved... |
| 11:36 | <jgraham> | Presumably the typical "look but don't touch" licenses |
| 12:41 | <jgraham> | <!doctype html><form><table></form><form></table></form> |
| 12:41 | <jgraham> | Is it me or is that broken with the changes to make <form> in <table> set the form pointer |
| 12:42 | <jgraham> | or it could just be bad wording in the spec |
| 12:42 | <jgraham> | But when you hit the final <form> there is *a* form element in scope; the first one |
| 12:43 | <jgraham> | But that is not the one the form element pointer points to |
| 12:43 | <jgraham> | so when you try to remove the one the form element pointer points to, badness happens |
| 12:45 | <jgraham> | The end tag ... form stuff says "the stack of open elements does not have node in scope", but the scope stuff only deals with elements of a specific type (presumably meaning (ns, name)), not specific nodes |
| 12:46 | <jgraham> | No wait I am misreading |
| 12:47 | <jgraham> | OK so the problem is that _almost_always_ you just look for an element with a specific name |
| 12:47 | <jgraham> | But the algorithm is wordded so that you are always looking for a specific ID |
| 12:48 | <jgraham> | s/id/node/ |
| 13:02 | <annevk5> | hmm, not much spec updates but email is now below 400 |
| 13:03 | <annevk5> | might be lowest point in 3-4 years |
| 13:04 | <Ms2ger> | It was at zero last October, wasn't it? |
| 13:04 | <annevk5> | that seems unlikely |
| 13:05 | <annevk5> | I have email from 2006 here :) |
| 13:05 | <Philip`> | Which spec do you mean? |
| 13:05 | <Ms2ger> | Oh, you |
| 13:06 | <jgraham> | Wasn;t it 0 excluding datagrid or something |
| 13:06 | <annevk5> | Philip`, various; cssom; cssom view; xhr; xhr2; etc. |
| 13:07 | <annevk5> | yeah, http://www.whatwg.org/issues/data.html?period=6 is slightly worse |
| 13:09 | <jgraham> | "Estimated date for last e-mail based on the data above: 2006-02-02" |
| 13:09 | <annevk5> | it says 2013-03-20 here |
| 13:09 | <jgraham> | Set it to period=12 |
| 13:10 | <annevk5> | when set to max it gives 2011-07-19 |
| 13:10 | <annevk5> | heh |
| 13:10 | <jgraham> | Hint taht your estimation algorithm is bad: it depends on the amount of data you are displaying :) |
| 13:11 | <annevk5> | to be fair it does state that |
| 13:11 | <Philip`> | It's not like it'd be less wildly inaccurate and meaningless if it used more of the data |
| 13:11 | <matjas> | Is there such a thing as XHTML5 + RDFa? I found http://dev.w3.org/html5/rdfa/rdfa-module.html — does this mean it’s just not available yet? |
| 13:11 | <jgraham> | Well only if you consider "the data above" to be the range that you happen to be displaying |
| 13:12 | <jgraham> | rather than the full dataset |
| 13:14 | gsnedders | wants a sing-along (heck, just a DVD would be a start) copy of Avenue Q :D :D :D |
| 13:15 | <matjas> | Re: XHTML5 + RDFa — I'm asking because a client of mine explicitly asked for XHTML + RDFa. Do I have to use old-school XHTML 1 + RDFa or is there a way for me to use (X)HTML5 here? |
| 13:15 | jgraham | wants to never hear the words "singalong" and "gsnedders" in the same sentence without the words "will be forbidden from" |
| 13:15 | <gsnedders> | jgraham: Aww, even though you missed the Sound of Music sing-along at the weekend? |
| 13:15 | <jgraham> | "missed" |
| 13:16 | <gsnedders> | It was awesome. |
| 13:17 | <jgraham> | s/esome/ful/ |
| 13:17 | <annevk5> | matjas, dunno, we don't do RDFa |
| 13:17 | <annevk5> | matjas, tell him about Microdata |
| 13:17 | <annevk5> | or her |
| 13:17 | <gsnedders> | or them |
| 13:18 | <annevk5> | or everyone really |
| 13:18 | <annevk5> | up to you ;p |
| 13:18 | <matjas> | annevk5: sadly, there's no convincing him. His company provides "XML and RDFa services"… |
| 13:19 | <jgraham> | So what we're saying is because you asked a question about RDFa you have to become a Microdata evangelist now |
| 13:19 | <jgraham> | to atone for your sins |
| 13:19 | jgraham | almost types sines |
| 13:19 | <matjas> | he just asked for "XHTML+RDFa"; I'm sure he's not gonna make any use of RDFa though. |
| 13:19 | <annevk5> | matjas, there was this article on ALA about nicely rejecting clients ;p |
| 13:19 | <annevk5> | matjas, but euh, there some proposal for HTML5+RDFa |
| 13:20 | <matjas> | annevk5: haha. wish i was popular enough to just do that :) |
| 13:20 | <matjas> | annevk5: this? http://dev.w3.org/html5/rdfa/rdfa-module.html |
| 13:20 | <annevk5> | http://www.w3.org/TR/rdfa-in-html/ |
| 13:21 | <annevk5> | http://dev.w3.org/html5/rdfa/ is the editor's draft |
| 13:23 | <matjas> | Hmm, so atm <html version="HTML+RDFa 1.1" lang="en"> is the way to go. Sadly the HTML5 spec considers this an error, since @version is obsolete. :'( |
| 13:25 | <jgraham> | You could probably invoke the "other applicable specifications" clause |
| 13:26 | <jgraham> | Since I think they made an @version specification |
| 13:29 | <matjas> | I guess I’ll have to stick to XHTML1+RDFa then. *cries* |
| 13:29 | <matjas> | I’d love to tell my client he doesn’t really need it, but that would be the equivalent of telling him to start a new business |
| 13:29 | <matjas> | DO NOT WANT |
| 13:34 | <annevk5> | you can map RDF to Microdata |
| 13:34 | <annevk5> | why not do that? |
| 13:54 | <matjas> | annevk5: heh, wut? got an example of that? |
| 14:00 | <Ms2ger> | matjas, http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#examples-4 is one example |
| 14:01 | <matjas> | Thanks |
| 14:02 | <annevk5> | so Ms2ger, wdyt, attempt to remove lowsrc as well or keep it? Firefox has the ability to figure it out, but I'm not sure how much they like being guinea pigs :) |
| 14:02 | annevk5 | wonders if sicking is around |
| 14:03 | <Ms2ger> | You noticed I broke Chatzilla by killing document.height? :) |
| 14:03 | <annevk5> | oh god |
| 14:03 | <annevk5> | no i didn't |
| 14:41 | <boblet> | hey all, I wrote some stuff about microformats in HTML5 http://html5doctor.com/microformats/ |
| 14:56 | <annevk5> | hmm, forgot to get back to my WAI-ARIA comments |
| 14:57 | <annevk5> | guess it's too late now :/ |
| 15:02 | <annevk5> | it's still hopelessly complicated too |
| 15:03 | <annevk5> | e.g. http://www.w3.org/WAI/PF/aria/20100616/roles#textalternativecomputation |
| 15:03 | <annevk5> | would be interesting to see the test suite for that section |
| 15:37 | <annevk5> | whoa, that took long |
| 15:37 | <annevk5> | http://lists.w3.org/Archives/Public/public-pfwg-comments/2010JulSep/0023.html |
| 15:56 | <cheeser> | how close are we to a stable-ish websocket spec? |
| 15:56 | <cheeser> | trying to decide if should update to the latest spec now or just wait a bit. |
| 15:56 | <gsnedders> | Good question. |
| 15:56 | <cheeser> | it's likely the browsers haven't caught up anyway |
| 15:58 | <annevk5> | cheeser, about as close as yesterday or a couple of weeks ago |
| 15:59 | <cheeser> | yeah. that's what i'm afraid of. ;) |
| 15:59 | <cheeser> | i think the only real change i'm missing is the frame formatting |
| 16:00 | <cheeser> | which, i have to say, i'm excited about. makes frame parsing much simpler. |
| 16:00 | <jgraham> | cheeser: The editor changed and the new one wants something much more complex |
| 16:00 | <jgraham> | Unless that is what you mean |
| 16:01 | <cheeser> | i mean the change that made all frame length delimited. |
| 16:01 | <cheeser> | is Hixie no longer the editor? |
| 16:01 | <jgraham> | He is editing the API |
| 16:01 | <jgraham> | Ian Fette is editing the protocol |
| 16:01 | <jgraham> | He wants something way more complex |
| 16:01 | <cheeser> | ah |
| 16:01 | <cheeser> | any url on that? or a mailing list? |
| 16:02 | <jgraham> | looking.... |
| 16:02 | <cheeser> | looks like maybe thehybi list |
| 16:02 | <smaug____> | was it actually announced somewhere that I.Fette is the new editor |
| 16:02 | smaug____ | needs to read all the hybi emails |
| 16:03 | <jgraham> | smaug____: Yeah, on the list |
| 16:03 | <cheeser> | i'm afraid to sign for that list. it's not *entirely* job related excpet for the websockets bits. |
| 16:03 | <cheeser> | so what's he want? |
| 16:03 | <jgraham> | http://www.ietf.org/mail-archive/web/hybi/current/msg03154.html |
| 16:03 | <jgraham> | cheeser: It is all WebSockets related now |
| 16:03 | <jgraham> | But it is fairly high traffic |
| 16:04 | <jgraham> | My long term average is to read about 30% of the mail |
| 16:05 | <cheeser> | ouch |
| 16:05 | <jre> | cheeser, if your job has to do with websockets, it seems really strange if you aren't following the mailing list where the protocol is being developed. |
| 16:05 | <cheeser> | i'm not entirely sure how to read that "image" |
| 16:05 | <cheeser> | wth? |
| 16:05 | <cheeser> | wtf was that? |
| 16:05 | <cheeser> | :) |
| 16:05 | <jgraham> | That was Julian I guess |
| 16:06 | <jgraham> | Dunno why he joined to say that one thing |
| 16:06 | <annevk5> | that's his style |
| 16:06 | <cheeser> | or how he even knew about it |
| 16:06 | <jgraham> | At least we know who reads the logs now :) |
| 16:06 | <annevk5> | irc logs |
| 16:06 | <cheeser> | heh |
| 16:06 | <annevk5> | jgraham, that's nothing new |
| 16:06 | <jgraham> | annevk5: I knjow :) |
| 16:07 | <annevk5> | though i wonder if he runs some kind of script as refreshing it all the time seems rather time consuming |
| 16:07 | <jgraham> | cheeser: Anyway, the diagram is, I think just bits across the top |
| 16:07 | <jgraham> | split into 32 bit lines |
| 16:08 | <cheeser> | yeah. i was noticing that. so the payload length is either 7bits or 64bits long if the frame-length == 127 |
| 16:09 | <jgraham> | Something like that |
| 16:10 | <jgraham> | annevk5: Maybe he has a script that outputs it to his irc window |
| 16:12 | <cheeser> | i should subscribe, though. i'll just have to manage the volume of it. i end up skimming enough of the dev lists i'm already on. and if it is mostly WS these days, it can't hurt. |
| 16:12 | <cheeser> | are these framing questions also a subset of hybi then? |
| 16:13 | <jgraham> | In what way? |
| 16:13 | <jgraham> | They are about WebSockets |
| 16:13 | <cheeser> | i'm not sure what the relationship between websockets and hybi is. |
| 16:13 | <jgraham> | The HyBi list is currently focused entirely on WebSockets |
| 16:14 | <jgraham> | It is of course not impossible that that will change in the future |
| 16:14 | <cheeser> | seems like WS is a bidi addition on top of HTTP while hybi is really looking at replacing all of HTTP with something a bit more modern and bidi in general |
| 16:14 | <jgraham> | You might be thinking od spdy? That's not really related |
| 16:15 | <jgraham> | WebSockets is only really tangentially related to HTTP |
| 16:15 | <jgraham> | It's just 2 way comunication that works with the requirements of browsers |
| 16:15 | <cheeser> | right. i get the websocket/http distinction. |
| 16:15 | <jgraham> | So hybi is a bit of a misnomer |
| 16:16 | <cheeser> | isn't hybi a full protocol spec? |
| 16:16 | <cheeser> | or more of just an umbrella project? |
| 16:16 | <jgraham> | hybi doesn't really exit |
| 16:16 | <jgraham> | *exist |
| 16:16 | <jgraham> | it is just the name of the mailing list |
| 16:16 | <cheeser> | oh, i see. |
| 16:16 | <jgraham> | not a concrete technology |
| 16:16 | <cheeser> | i thought it was an actuall proposed spec |
| 16:17 | <cheeser> | i see. and spdy then is the actuall "let's get rid of http and use this" tech then. yes? |
| 16:19 | <cheeser> | so how do I subscribe to that mailing list? not seeing anything in the messages in the archive or on the ietf site |
| 16:19 | <jgraham> | https://www.ietf.org/mailman/listinfo/hybi |
| 16:20 | <cheeser> | ah. just found that. thanks. |
| 16:20 | <olivvv> | hi. quick question, is there in the whole package to come with html5, things that allow to communicate with locally installed apps ? |
| 16:21 | <olivvv> | its possible in with flash, java applets |
| 16:22 | <jgraham> | olivvv: No. What are you trying to achieve? |
| 16:23 | <cheeser> | there. signed up. |
| 16:24 | <cheeser> | how stable is spdy? is that on topic here? i've been eyeballing implementing that in glassfish/grizzly as well. |
| 16:24 | <olivvv> | jgraham:Computer - telephony Integration |
| 16:25 | <olivvv> | jgraham:its for a call center, using a small soft, the clients can provided cc number through the phone without the operator to see it |
| 16:26 | <olivvv> | this software needs then to communicate with the webapp in the browser |
| 16:27 | <olivvv> | anyone ? |
| 16:28 | <jgraham> | cheeser: There is not much that is off-topic here (see the /topic) but I don;t think most people follow spdy closely |
| 16:28 | cheeser | nods. |
| 16:29 | <jgraham> | olivvv: I'm not sure that is possible. It at least sounds like a non-Web use case which is typically not such a high priority |
| 16:29 | <cheeser> | spdy seems mostly a google effort at this point but it's certainly intriguing |
| 16:29 | <jgraham> | olivvv: Although maybe it could work with <device> or so somehow |
| 16:30 | <olivvv> | jgraham: I know it is strange, but since flash and applets are able to do that, I thought somebody would have thought of a standard alternative |
| 16:31 | <jgraham> | olivvv: Communicating with local devices is generally hard because there are all sorts of security problems |
| 16:31 | <jgraham> | But the <device> element is under development for e.g. video cameras |
| 16:32 | <olivvv> | jgraham:thx I ll google that |
| 16:50 | <sean`> | up |
| 17:33 | <AryehGregor> | Yay progress: http://www.imperialviolet.org/2010/08/16/dnssectls.html |
| 17:40 | annevk5 | would not mind getting rid of CAs if that was possible |
| 17:49 | <annevk5> | that blog is pretty interesting |
| 17:49 | <annevk5> | http://www.imperialviolet.org/2010/06/25/overclocking-ssl.html |
| 17:50 | <annevk5> | apparently Google Chrome is using SPDY to some extent already? |
| 17:51 | <cheeser> | seems like it supports it, yeah. |
| 17:51 | <cheeser> | i'm not aware of any servers that do, though. |
| 17:52 | <annevk5> | it says Google servers do |
| 17:52 | <cheeser> | what do they run, do you know? |
| 17:52 | <cheeser> | not stock anything i'm aware of. |
| 18:33 | <AryehGregor> | whatwg.org is kind of flaky sometimes. Maybe it should get better hosting. |
| 18:34 | <AryehGregor> | (w3.org isn't perfect either, but it seems to be more reliable, which makes us look bad) |
| 18:49 | <Hixie> | AryehGregor: there's some problem that's been affecting stability recently. hardware or something. not sure what, still studying it. |
| 18:50 | gsnedders | is innocent… this time. |
| 18:51 | <AryehGregor> | The ellipsis character looks very annoying in fixed-width fonts. |
| 18:51 | <Hixie> | gsnedders: did you see my comment about bug 10002? (i think it was 10002) |
| 18:51 | AryehGregor | doesn't know why it exists, it's usually much narrower than an ellipsis is supposed to be, even in proportional fonts |
| 18:51 | <gsnedders> | Hixie: See what myself and MikeSmithX discussed earlier |
| 18:51 | <Hixie> | k |
| 18:51 | <Hixie> | in #whatwg? |
| 18:51 | <gsnedders> | Yeah |
| 18:54 | <Hixie> | got it |
| 19:04 | <Hixie> | oh cool, dreamhost replied to my extremely vague request. They've done something to fix the problem and they'll move us to another server if the problems continue anyway. |
| 19:06 | <AryehGregor> | Maybe you should upgrade to dedicated hosting. |
| 19:06 | <Hixie> | sure, send me the money to fund it and i'll get right on that :-) |
| 19:07 | <AryehGregor> | Is dedicated hosting really so much more expensive? |
| 19:07 | <Hixie> | right now the dreamhost hosting is essentially free |
| 19:07 | <AryehGregor> | How does that work out? |
| 19:07 | <Hixie> | the money i get back on referrals is higher than the cost of my account |
| 19:08 | <Hixie> | and my account has unlimited everything -- bandwidth, disk, domains, users, etc |
| 19:08 | <Hixie> | because i've been a customer for so long |
| 19:08 | <Mukke|home> | i wonna implement html 5, where to start ? |
| 19:08 | AryehGregor | cringes at a VPS provider claiming to provide "unlimited" anything |
| 19:08 | <Hixie> | AryehGregor: "unlimited" just means they don't charge me by the byte or whatnot |
| 19:08 | <AryehGregor> | Ah. |
| 19:09 | <Hixie> | if i run out of disk, they move me to a machine with more disk |
| 19:09 | <Hixie> | etc |
| 19:10 | <AryehGregor> | I say all this as someone whose personal website is hosted on a machine I happen to run, where the main site costs $500/month for bandwidth alone, so I haven't really had any reason to look into the lower end of the market. |
| 19:10 | <Hixie> | for the last 6 months of hosting several dozen domains and subdomains i've paid $21 |
| 19:11 | <AryehGregor> | Mukke|home, you should start by saying "What features do I want to add that my users will want?", and go from there. If you just want to be able to claim that your pages are HTML5, change the doctype to <!doctype html>, run it through validator.nu, and fix all the errors it reports. |
| 19:11 | <AryehGregor> | This is probably not worth the effort, though. |
| 19:11 | <Mukke|home> | yeh |
| 19:12 | <AryehGregor> | Hixie, well, I've paid $0, since I don't pay the bills for my server. I just freeload. :) |
| 19:12 | <Hixie> | :-) |
| 19:12 | <AryehGregor> | In exchange for having fun administering the machine as root. |
| 19:12 | <AryehGregor> | Like fixing it when we have downtime. |
| 19:12 | <Hixie> | i have root too, but don't have to admin it :-) |
| 19:12 | <AryehGregor> | Yeah, that's the disadvantage of colocation, you have to handle everything. |
| 19:12 | <AryehGregor> | Down to failed hardware. |
| 19:13 | <AryehGregor> | It sure saves on cost, though. |
| 19:13 | <Mukke|home> | AryehGregor, i never have written html 4 like i should have, i wanno do it right and with proper coding in html 5 now |
| 19:13 | <AryehGregor> | Mukke|home, you can use http://validator.nu/ as a guide to see if your website is good HTML5. You might also look at: http://dev.w3.org/html5/html-author/ |
| 19:16 | <Mukke|home> | AryehGregor, i need a special featere and i'm almost sure it doesn't excists that whyi came here aswell |
| 19:16 | <AryehGregor> | Then ask. |
| 19:16 | <Mukke|home> | can i shape a div hexagon ? |
| 19:17 | <Mukke|home> | or any elemnt at all that can have a background_image |
| 19:17 | <AryehGregor> | No, all elements in CSS are boxes. |
| 19:17 | <AryehGregor> | This might be a question for #css, though. |
| 19:58 | <AryehGregor> | "Case canonicalization for reflected enumerated attributes limited to known values" |
| 19:58 | <AryehGregor> | I like it when you can write massive strings of gibberish like that and they're actually the only reasonable way to refer to something, not gratuitous at all. |
| 19:59 | <AryehGregor> | My favorite math theorem name is "Parametrized Transfinite Recursion Theorem". |
| 20:04 | <gsnedders> | Where the verb there? |
| 20:05 | <AryehGregor> | Where's the verb where? |
| 20:05 | <gsnedders> | Indeed. |
| 20:05 | <gsnedders> | :) |
| 20:05 | <gsnedders> | (Freudian slip?) |
| 20:05 | <AryehGregor> | Hmm, wait, this is weirder than I thought. |
| 20:06 | <AryehGregor> | Nobody seems to be properly implementing IDL gets at all on reflected enumerated attributes limited to known values. |
| 20:06 | <AryehGregor> | Except maybe IE. |
| 20:06 | AryehGregor | admits to copy-pasting the phrase "reflected enumerated attributes limited to known values" |
| 20:06 | <cheeser> | ugh. CORBA flashbacks. |
| 20:09 | <AryehGregor> | I think I pressed Ctrl-Shift-W or something by mistake. |
| 20:09 | <AryehGregor> | Whoever decided to make an easy-to-press way to quit in any application, ever, deserves to be smacked on the head. |
| 20:09 | <Hixie> | i hit command-q instead of command-w all the time |
| 20:09 | <Hixie> | drives me crazy |
| 20:09 | <AryehGregor> | Good thing I have a shell script that positions my XChat windows exactly as I want them on startup. |
| 20:09 | <cheeser> | i love how w and q are right next to each other so it's easy to do cmd-q instead. |
| 20:10 | Hixie | like his dvorak keyboard for that reason :-) |
| 20:10 | <AryehGregor> | One of my first commits to MediaWiki was to remove the keyboard shortcut for "log out". |
| 20:10 | <gsnedders> | Run stuff in a screen so you can just reattach your terminal to it after restarting your terminal? :P |
| 20:11 | <AryehGregor> | It doesn't happen for me with terminals. |
| 20:11 | <gsnedders> | (That does seriously work quite well for me) |
| 20:11 | <cheeser> | screen+irssi ftw |
| 20:11 | <gsnedders> | screen+* ftw. |
| 20:11 | <AryehGregor> | The only risk is that Ctrl-Shift-uXXXX is used for Unicode characters, and the X's sometimes need to be E, which is right next to W, and Ctrl-Shift-W exits. |
| 20:11 | <AryehGregor> | For some bizarre reason. |
| 20:11 | <AryehGregor> | I actually never learned how to use screen. |
| 20:11 | <AryehGregor> | Never needed to yet. |
| 20:12 | <gsnedders> | AryehGregor: That's what you say before you use it. |
| 20:12 | <AryehGregor> | What else does it do? |
| 20:12 | <Hixie> | screen(1) is fantastic. |
| 20:12 | <Hixie> | it makes the mapping of terminal client to terminal be many-to-many |
| 20:12 | <Hixie> | so you can just reconnect to an open session |
| 20:13 | <Hixie> | or have multiple open sessions |
| 20:13 | <Hixie> | or whatever |
| 20:13 | <cheeser> | you can share screens remotely ;) |
| 20:13 | <gsnedders> | And then you can hit ctrl+c when someone else is at a password prompt! |
| 20:13 | <gsnedders> | And then before they realize it, you've got their password! |
| 20:13 | <cheeser> | or press space randomly since they can't see the nonechoed input anyway. |
| 20:14 | <gsnedders> | indeed |
| 20:14 | <Hixie> | my favourite use: i was ssh'ed into a session on a remote host from a laptop. My partner's cat walks over and sits on my keyboard to get my attention. So I drag out my other laptop and just ssh in -- yay screen, i could pick up where i left off. |
| 20:14 | <Hixie> | the cat was quite miffed. |
| 20:14 | <cheeser> | heh |
| 20:15 | <AryehGregor> | Okay, I'm sold. |
| 20:15 | <AryehGregor> | I'll have to learn how to use it. |
| 20:15 | <gsnedders> | Hixie: How many cats do you have around you now? :P |
| 20:15 | <Hixie> | one and three quarters. |
| 20:15 | <Hixie> | actually the three quarters is outside |
| 20:15 | <Hixie> | so one |
| 20:15 | <gsnedders> | Well, not at this very moment in time. |
| 20:15 | <Hixie> | AryehGregor: put screen -a -A -xRR -U -e^Pp in your .bash_profile |
| 20:15 | <gsnedders> | But in general |
| 20:16 | <gsnedders> | But how many cats does your partner have? |
| 20:16 | gsnedders | original wrote that with cats and partner the other way around… uh… |
| 20:16 | <Hixie> | AryehGregor: then hit ^P ^C to open a new terminal, ^P ^N and ^P ^H to go to the next/previous terminal, ^P ^P to toggle with the last terminal, and ^P 0-9 to jump to the 0-9th terminal |
| 20:17 | <gsnedders> | ^P " to list all termianls |
| 20:17 | <gsnedders> | *terminals |
| 20:17 | <AryehGregor> | Why ^P instead of the default ^A? |
| 20:17 | <Hixie> | because i use ^A to go to the start of the line |
| 20:17 | <AryehGregor> | I use Home for that, since I only use desktops. |
| 20:18 | <Hixie> | oh and ^P ? to get help if you need it |
| 20:18 | <Hixie> | if you don't use ^A then feel free to remove the -e^Pp part of the command line :-) |
| 20:19 | <AryehGregor> | Is there any reason to use screen to manage multiple terminals, instead of just using tabs? |
| 20:19 | <AryehGregor> | Also, does anything bad happen now if I close my terminal without typing "exit" or such? |
| 20:19 | <AryehGregor> | (like lots of screens accumulating forever?) |
| 20:20 | <Hixie> | you can use tabs too |
| 20:20 | <Hixie> | do both |
| 20:20 | <Hixie> | that way if you need to connect from a single tab, you can still do that and use all the screens |
| 20:20 | <Hixie> | but you can still connect multiple times and have each tab be a different screen in screen(1) |
| 20:22 | <Hixie> | (that's what all the "-a -A -xRR -U" stuff does -- it makes it so there's one session and you can connect as many times as you want) |
| 20:22 | <AryehGregor> | Wait, so now if I connect from multiple tabs, I'll get the same screen in all of them, right? So I'll want to do ^A^C to make some new ones, or occupy my existing ones once I've got them established. |
| 20:23 | <AryehGregor> | Okay, let me fiddle with this now. |
| 20:23 | <Hixie> | yeah, by default with those arguments each instance will connect to the same screen, but then you can change it to be whichever you want |
| 20:23 | <Hixie> | basically it's like each tab can see all the other tabs |
| 20:23 | <AryehGregor> | Why does it say "Attaching from inside of screen?" every time I connect? |
| 20:24 | <Hixie> | if you put it in .bash_profile, then opening a shell connects already |
| 20:24 | <AryehGregor> | Yeah, I did. |
| 20:25 | <Hixie> | gotta go, bbiab, sorry for leaving you with a weird setup :-) you'll figure it out :-) see man screen ! |
| 20:25 | <gsnedders> | Or ask here :) |
| 20:25 | <gsnedders> | You get that message every time you connect to a screen from within sceen |
| 20:26 | <AryehGregor> | What does "from within screen" mean? |
| 20:26 | AryehGregor | reads man screen |
| 20:26 | <gsnedders> | Like, when you're in a terminal in a screen window |
| 20:27 | <cheeser> | recursive screen connects |
| 20:27 | <gsnedders> | Basically: you probably don't want screen running within screen |
| 20:27 | <AryehGregor> | I just put the command Hixie told me to into my .bash_profile or equivalent. |
| 20:32 | <jgraham> | AryehGregor: (if you are using Ubuntu it comes wih some utility to make fancy .screen files) |
| 20:32 | <AryehGregor> | I am. |
| 20:32 | <jgraham> | screen-profile or something. I think it is supposed to run the first time you run screen though |
| 20:32 | <jgraham> | So if it didn't I don't know how to fix it |
| 20:33 | <jgraham> | Because I tried installing it and it didn't work... |
| 20:33 | <jgraham> | (after I was already using screen) |
| 20:40 | AryehGregor | finds screen on Solaris behaves differently, adds a conditional to his .bash_profile to avoid it |
| 20:41 | <AryehGregor> | Oh, hmm. |
| 20:42 | <AryehGregor> | Maybe what happens is that I log in, it starts a screen, the screen starts bash, bash tries to start screen, . . . |
| 20:42 | AryehGregor | got a whole procession of screens on Solaris, had to hold down Ctrl-D to kill them all |
| 20:50 | AryehGregor | uses if ! [ "$TERM" = 'screen' ]; then ... |
| 20:51 | <AryehGregor> | Now I feel like I'm living in the Matrix. I have this magical screen thing, and if I do Ctrl-A d I get back to real life temporarily. |
| 20:52 | <cheeser> | that's so 20th century. Inception is the modern meme. ;) |
| 20:54 | <AryehGregor> | Well, the Matrix movies were practically the last movies I ever watched, so I probably won't watch that. |
| 20:54 | <cheeser> | haha. nice. |
| 20:57 | <gsnedders> | Inception was pretty good |
| 21:01 | <AryehGregor> | Okay, "man screen" is absurdly long. |
| 21:02 | <AryehGregor> | $ man screen | wc |
| 21:02 | <AryehGregor> | 2566 21969 152831 |
| 21:02 | <gsnedders> | Uh, yeah. I've never read it all. |
| 21:02 | <AryehGregor> | No way am I reading that whole thing. |
| 21:02 | <AryehGregor> | I'll work it out as I go along. |
| 21:02 | <gsnedders> | Irn-bru cocktails? Oh dear… |
| 21:02 | <gsnedders> | Sometimes things on Facebook scare me |
| 21:07 | <hober> | put "nethack on" in .screenrc for extra awesomeness |
| 21:11 | <hober> | also, "term screen-256color" is great assuming the terminal app you use supports 256 colors (sadly Mac OS X's Terminal.app doesn't) |
| 22:17 | <annevk5> | gmail down? |
| 22:19 | <sean`> | nope |
| 22:19 | <sean`> | guess they don't want you there anymore |
| 22:20 | <annevk5> | gar |
| 22:27 | <annevk5> | getting 502s |
| 22:33 | <llrcombs> | OK, http://www.w3.org/TR/html5/timers.html#dom-navigator-registercontenthandler has a couple flaws I'd like to point out |
| 22:33 | <llrcombs> | 1. no explanation on how/if registerContentHandler can cause local resources to be opened in a web app |
| 22:34 | <llrcombs> | 2. no explanation/example for what exactly %s is replaced with in registerProtocolHandler |
| 22:37 | <gsnedders> | 1. Surely this follows from the inability for local files to be read by sites? What it applies to seems like intrensically impl-specific behaviour. |
| 22:37 | <gsnedders> | 2. The definition of the url argument seems to have that |
| 22:37 | <llrcombs> | well, then it should be explicitly defined in the spec |
| 22:38 | <llrcombs> | and the url param spec doesn't define if it's supposed to include the protocol name, etc |
| 22:38 | <annevk5> | http://www.boingboing.net/2010/08/17/is-the-web-really-de.html -- lol wtf did Wired do |
| 22:39 | <annevk5> | llrcombs, you can file bugs |
| 22:39 | <annevk5> | llrcombs, or email |
| 22:39 | <gsnedders> | llrcombs: In terms of conformance requirements for calls to it? |
| 22:39 | <gsnedders> | llrcombs: The UA behaviour is clearly defined |
| 22:39 | <gsnedders> | llrcombs: 'When the user agent uses this URL, it must replace the first occurrence of the exact literal string "%s" with an escaped version of the absolute URL of the content in question (as defined below), then resolve the resulting URL, relative to the base URL of the entry script at the time the registerContentHandler() or registerProtocolHandler() methods were invoked…' |
| 22:40 | <gsnedders> | (follow the xref for resolve, and you'll find out it will work with any relative URL) |
| 22:41 | <llrcombs> | so that's 2 |
| 22:41 | <llrcombs> | alrighty |
| 22:41 | <llrcombs> | upon first reading the first bit of this spec, I thought it might send a POST request with the file's contents |
| 22:42 | <gsnedders> | llrcombs: for 1? |
| 22:42 | <llrcombs> | yeah |
| 22:42 | <llrcombs> | then read a bit more |
| 22:42 | <gsnedders> | llrcombs: No, it's quite clearly a GET request |
| 22:42 | <llrcombs> | found it sent a GET request with a URL |
| 22:42 | <llrcombs> | yeah, that's clea |
| 22:42 | <llrcombs> | r |
| 22:42 | <llrcombs> | I just meant that was my thought when i read the first tiny bit of the spec |
| 22:42 | <llrcombs> | no issue in wording or anything |
| 22:43 | <gsnedders> | Yeah, there's plenty of stuff that relies rather heavily on cross-references |
| 22:43 | <llrcombs> | I was a bit confuzed when I saw it was a GET request, as I know web apps can't access file: URLs |
| 22:43 | <gsnedders> | But making it clear everywhere would probably double the length of the spec |
| 22:43 | <llrcombs> | yeah |
| 22:43 | <llrcombs> | I get that |
| 22:43 | <llrcombs> | I didn't say there was a clarity issue with the fact that it's a GET request with the URL |
| 22:58 | <AryehGregor> | I saw a "the browser is dead, apps are the future" article someplace else a while ago. The Atlantic, I think. |
| 22:58 | <AryehGregor> | All pretty silly stuff. |
| 22:58 | <AryehGregor> | (I hope, otherwise I'm wasting my time here) |
| 22:59 | <AryehGregor> | "The defenders of the unfettered Web have their hopes set on HTML5 — the latest version of Web-building code that offers applike flexibility — as an open way to satisfy the desire for quality of service." |
| 23:00 | <AryehGregor> | Dang right. |
| 23:08 | <AryehGregor> | I like the commenter who points out that according to Wired's graph, DNS died in about 1995. |
| 23:19 | <llrcombs> | loltru |