| 01:20 | <heycam> | Hixie, nope, no way to specify that |
| 01:21 | <Hixie> | how should i do it? |
| 01:22 | <heycam> | i do have to fix up arrays/sequences quite a bit, but maybe you could do sequence<sequence<Object>>? |
| 01:23 | <Hixie> | k |
| 03:40 | <karlcow> | [13:08] <cryzed> And I somehow think that using BeautifulSoup is easier than lxml |
| 03:40 | <karlcow> | Did you see http://blog.ianbicking.org/2008/12/10/lxml-an-underappreciated-web-scraping-library/ |
| 05:48 | <john_fallows> | Hixie: fyi, i just subscribed to hybi |
| 07:09 | <Hixie> | john_fallows: cool |
| 07:09 | <john_fallows> | just read through all the hybi archives *whew* |
| 07:10 | <Hixie> | heh |
| 07:14 | <john_fallows> | tell me if i understand rHTTP correctly - someone wanted to relax the client-initiated constraints of HTTP, so they used 101 Switching Protocols with Upgrade header, and then chose the same half-duplex protocol only in the opposite direction? |
| 07:21 | <john_fallows> | rHTTP seems to be missing the point, once you have completed the Upgrade, you could choose to leverage the full power of raw TCP at that point, letting you implement rHTTP or any other protocol |
| 07:23 | <john_fallows> | of course, JavaScript has no binary type (yet) so you would need minimal framing to determine raw binary (just like TCP) versus text - giving us WebSocket as currently specified |
| 07:24 | <Hixie> | you've pretty much summed up what appears to be the current state of affairs |
| 07:25 | <Hixie> | there also seems to be a lack of understanding of the web browser security model |
| 07:26 | <Hixie> | which doesn't help |
| 07:26 | <Hixie> | but anyway |
| 07:26 | <john_fallows> | that's a good point, and critically important in understanding where WebSocket came from |
| 07:27 | <john_fallows> | there was also some discussion about formalizing how proxies should behave in the presence of an Upgrade header |
| 07:28 | <john_fallows> | i thought that the Upgrade header was hop-by-hop by design, so it wouldn't be appropriate to give it any meaning beyond the current hop - necessitating CONNECT instead, just like https, right? |
| 07:31 | <john_fallows> | or is the concern more about transparent HTTP proxies that would not receive a CONNECT ? |
| 07:31 | <Hixie> | i'm not an expert when it comes to proxies |
| 07:32 | <john_fallows> | in any case, it's probably a good idea for the IETF to formalize how proxies (transparent or otherwise) should behave in the presence of a 101 Switching Protocols response |
| 07:36 | <Hixie> | mcarter and i had dinner the other day and discussed some of the feedback, btw; in particular around framing, and other ways to do the handshake |
| 07:37 | <Hixie> | i think we concluded that there were some options but none especially compelling |
| 07:37 | <Hixie> | in particular mcarter had some ideas around faking the SSL handshake instead of the HTTP one |
| 07:37 | <Hixie> | (faking a zero-encryption handshake) |
| 07:39 | <Hixie> | but that doesn't really work well when you think about it, sadly |
| 07:40 | <john_fallows> | i think i've also sent similar feedback to the public mailing list and you pointed out that it might introduce security concerns - which i agree with |
| 07:41 | <john_fallows> | btw, there are at least 2 ways to fake a zero encryption handshake |
| 07:41 | <Hixie> | oh? |
| 07:41 | <john_fallows> | 1) negotiate null encryption via SSL |
| 07:42 | <john_fallows> | 2) negotiate full encryption via SSL, and then use the subsequent "Upgraded" protocol to negotiate removal of the encryption |
| 07:42 | <Hixie> | true |
| 07:43 | <Hixie> | anyway. long and short of it is, i didn't make any changes |
| 07:43 | <Hixie> | not really sure what to do on hybi though |
| 07:43 | <john_fallows> | i think my previous feedback to the mailing list was #1, but #2 has more potential :) |
| 07:43 | <Hixie> | right now it's just me pushing a websocket-like approach |
| 07:44 | <Hixie> | and a bunch of others pushing elaborate multi-level protocols |
| 07:44 | <john_fallows> | well, i'll be around there to help others to understand your brilliance :) |
| 07:44 | <Hixie> | websocket is more mcarter's brilliance than mine at this point :-) |
| 07:45 | <Hixie> | ok time to go to bed |
| 07:45 | <Hixie> | nn |
| 07:45 | <john_fallows> | don't know if you realize this - but mcarter and i both designed a lot of his proposals together :-) |
| 07:46 | <john_fallows> | see you on hybi |
| 07:46 | <john_fallows> | nn |
| 08:02 | sayrer | reads @hybi and feels sorry for Hixie |
| 08:33 | <Philip`> | http://www.crummy.com/2009/04/09/0 - "I'll be writing an html5lib tree builder and packaging it and the lxml builder in Beautiful Soup for a while, but I think long-term the TreeBuilders should live with their parent projects." |
| 08:39 | <hsivonen> | I think the reception of Web Socket on hybi is sad but not surprising. |
| 08:57 | <hsivonen> | (it seems like the usual case of various people wanting to do very different things but getting bundled as one group because the may seem to be some commonality) |
| 09:03 | <hsivonen> | vipood suggests that I used several DNS servers to check the IP address of their site to restore my confidence until they restore the EV cert |
| 09:09 | <MikeSmith> | hsivonen: so what was the problem? |
| 09:36 | <hsivonen> | MikeSmith: a "hiccup" |
| 09:37 | <MikeSmith> | aka "fuckup" (I'd guess) |
| 09:42 | <hsivonen> | Hixie: did you happen to consider data-* elements, too when you added data-* attributes to the spec? |
| 10:04 | <MikeSmith> | hsivonen: I just tested and checked in the schema/assertions changes for <keygen> |
| 10:04 | <MikeSmith> | and 12+12 new tests (html+xhtml) |
| 10:05 | <MikeSmith> | writing the tests actually helped me catch a couple things I'd forgottent to add to the assertions.sch file |
| 10:07 | <MikeSmith> | so I guess I should complain less about making tests being a PITA |
| 10:08 | <hsivonen> | MikeSmith: nice! |
| 10:10 | <hsivonen> | Hixie: context for data-* element question: http://intertwingly.net/blog/2009/04/08/HTML-Reunification#c1239352657 |
| 10:29 | <hsivonen> | http://lists.w3.org/Archives/Public/public-xhtml2/2009Apr/0015.html |
| 10:35 | <MikeSmith> | hsivonen: lemme please know when you got latest v.nu deployed |
| 10:40 | <hsivonen> | MikeSmith: ok. I'm currently at an IP address that isn't allowed to trigger deployment |
| 10:40 | <hsivonen> | http://twitter.com/bradneuberg/status/1478859862 |
| 10:43 | <hsivonen> | http://steve-yegge.blogspot.com/2009/04/have-you-ever-legalized-marijuana.html |
| 10:43 | <hsivonen> | the spec item list in there is illustrative of many drive-by Web spec enhancement requests |
| 10:44 | <hsivonen> | particularly the point starting "WHAT THE EVER-LOVING *FUCK* ARE YOU PEOPLE SMOKING?" |
| 10:44 | <takkaria> | oh, I just ordered that book |
| 10:45 | <MikeSmith> | which "canvas shim" is Brad referring to, I wonder |
| 10:45 | <MikeSmith> | hsivonen: I like the "I personally do this all the time, except that I'm usually the ugly one." part |
| 10:48 | MikeSmith | scrolls to find the parts about smoking marijuana |
| 10:49 | <MikeSmith> | "I have smoked marijuana (and inhaled it, deeply) on more occasions than I can count. And yet I'm almost undoubtedly smarter than your kid that you're so goddamned worried about." |
| 10:49 | <MikeSmith> | well put |
| 10:54 | <MikeSmith> | hsivonen: have you read Fred Brooks's "The Mythical Man-Month"? |
| 10:57 | <hsivonen> | MikeSmith: yes |
| 10:58 | <hsivonen> | My boss at a previous job gave it to me when I left saying that I was a "productive individual" |
| 10:58 | <MikeSmith> | heh |
| 10:58 | <MikeSmith> | it's surprising to me how many people I've known in software product development who've not read that book |
| 10:58 | <MikeSmith> | this Steve Yegge post reminds me of it |
| 10:59 | <hsivonen> | s/previous/past/ I should say |
| 11:00 | <MikeSmith> | the decision-makers unfortunately don't read things like Steve's blog, nor Brooks |
| 11:00 | <MikeSmith> | they read instead stuff like "Who Moved My Cheese" |
| 11:00 | <MikeSmith> | or something like that |
| 11:01 | <MikeSmith> | books with like 10 words on each page, plus cartoons |
| 11:01 | <MikeSmith> | and then they distribute them to everybody in the company, because they think that stuff is brilliant |
| 11:02 | <MikeSmith> | I don't know how people get into those positions |
| 11:02 | <MikeSmith> | what the criteria are |
| 11:02 | <MikeSmith> | success certainly isn't a criterion in many cases |
| 11:03 | <MikeSmith> | I read today that the guy who's running Motorola's mobile business is the highest-paid software executive in the US |
| 11:03 | <MikeSmith> | despite the fact that part of their business is a massive failure |
| 11:07 | <MikeSmith> | the mobile part of their business being a failure, I mean |
| 11:18 | <takkaria> | good article, that |
| 11:18 | <MikeSmith> | aye |
| 11:19 | <takkaria> | I do work with a UK national drug law reform NGO and they're putting together proposals which try and answer most of his questions |
| 11:20 | <takkaria> | and yes, it is actually a pain in the arse to work out even how you go about decriminalising just possession well |
| 11:20 | <MikeSmith> | I posted a comment there under the alias "Winston" |
| 11:21 | <MikeSmith> | not really |
| 11:21 | <MikeSmith> | anyway, his "Because laws are pretty much like programs." premise is wrong |
| 11:21 | <MikeSmith> | as some of the comments point out |
| 11:22 | <MikeSmith> | but that doesn't mean most of his points aren't still valid |
| 11:26 | <takkaria> | the biggest obstacle to countries doing anything positive towards drug legalisation are the UN treaties on the matter, until they're gone there's a limit to what states can do in terms of a regulatory framework |
| 11:26 | <hsivonen> | MikeSmith: "Winston" overlooks the issue that stuff is way more complex than "my body is mine" in the case of distributed health care costs and the moral imperative not to just let sick people die |
| 11:26 | <MikeSmith> | hsivonen: yeah, true |
| 11:27 | <MikeSmith> | "my body is mine" is not the point |
| 11:27 | <MikeSmith> | plus, I think he made up the word "ayahuasca" |
| 11:28 | <MikeSmith> | ah, geez |
| 11:28 | <MikeSmith> | Wikipedia article on ayahuasca is great |
| 11:28 | <MikeSmith> | "any of various psychoactive infusions or decoctions prepared from the Banisteriopsis spp. vine, usually mixed with the leaves of the Psychotria bush" |
| 11:28 | <MikeSmith> | seems like even more made-up words |
| 11:30 | <MikeSmith> | hsivonen: obesity is a bigger problem that marijuana, as far as distributed health-care costs |
| 11:30 | <takkaria> | I think "And this is why Britain will continue to have a monarchy for centuries. The benefits of being a republic will never outweigh the effort of arging about what to put on postage stamps." is the best comment |
| 11:31 | MikeSmith | looks forward to the day when a fat marijuana bud appears on a postage stamp |
| 11:33 | <hsivonen> | MikeSmith: very likely. my point is that health care (and the investment in a person's education) make policy issues more complex than argumenting from pure individualism |
| 11:33 | <MikeSmith> | hsivonen: yeah, absolutely |
| 11:34 | <MikeSmith> | this is the key flaw in the so-called "libertarian" political philosophy |
| 11:35 | <MikeSmith> | I say "so-called" because that's what it's known by the last 50 years or so in the US at least |
| 11:35 | <MikeSmith> | but it's something quite different from what political libertarianism originally meant |
| 11:36 | <hsivonen> | the political vocabulary in the U.S. around liber* is quite transformed |
| 11:36 | <MikeSmith> | I don't mean the health-care issue specifically, but the general issue policy issues being more complex than pure individualism |
| 11:37 | <MikeSmith> | hsivonen: the political vocabulary of the US is like a pidgin language |
| 11:37 | <MikeSmith> | one of the few places that's worse is Japan, where there's no real debate at all |
| 11:37 | <MikeSmith> | nor no real political parties |
| 11:54 | <hsivonen> | http://m.alistapart.com/ what happened to CSS and media queries? |
| 12:00 | <hsivonen> | heh. jd worrying about the barrier for entry for new browsers: http://blogs.adobe.com/jd/2009/03/pervin_the_standards.html |
| 12:02 | <MikeSmith> | I bet jd is the kind of guy who spits a lot when he talks |
| 12:03 | <MikeSmith> | and I guarantee he does not smoke marijuana |
| 12:06 | <MikeSmith> | I wonder, if Adobe fired jd, what he would do |
| 12:12 | <gsnedders> | Is the barrier to entry not always going to get higher as more and more features are added? |
| 12:13 | <gsnedders> | So should we have never added the img element because it raised the barrier of entry? |
| 12:17 | hsivonen | mumbles about having to create a special Flash host process as a barrier to entry for Chrome |
| 12:55 | <roc> | I happen to agree with jd that CSS animations is a slightly dubious application of CSS |
| 12:55 | <roc> | I think that's about the only thing I agree with him about, though |
| 12:56 | gsnedders | posts a short comment, and will almost certainly not look at jd's blog for another year or so, and completely forget he ever published this comment |
| 14:13 | <hsivonen> | at the office now. Will redeploy v.nu |
| 18:05 | <gsnedders> | Philip`: Do you not have a working XML DOM Viewer? |
| 18:15 | <Philip`> | gsnedders: You mean like http://canvex.lazyilluminati.com/misc/dom-viewer/x.html ? |
| 18:54 | <gsnedders> | Philip`: Yes |
| 19:15 | <Philip`> | gsnedders: In that case, I don't have such a thing |
| 19:31 | <sayrer> | boy, webkit does not like hsivonen's downloadable font pages |
| 19:32 | <sayrer> | I thought Firefox was acting wonky |
| 19:34 | <hsivonen> | sayrer: just a flash of unfontified content or something worse? |
| 19:34 | <sayrer> | hsivonen, Firefox has a flash |
| 19:34 | <sayrer> | hsivonen, WebKit draws with no face at all |
| 19:34 | <sayrer> | Safari4 beta, I should say |
| 19:35 | <hsivonen> | sayrer: oh. that's bad. |
| 19:35 | <hsivonen> | sayrer: thanks |
| 19:35 | <sayrer> | hsivonen, once all the fonts are downloaded, it picks one |
| 19:35 | <hsivonen> | I thought the worst rendering issue was that the sans-serif font doesn't hint nicely in ClearType |
| 19:35 | <hsivonen> | it has been designed for real anti-aliasing |
| 19:39 | <sayrer> | http://people.mozilla.com/~sayrer/2009/04/hsivonen.png |
| 22:37 | <jwalden> | hsivonen: "sprikled" in the producing-xml page |
| 23:32 | <weinig_> | sayrer: out of curiosity, what would you expect hsivonen's producing-xml page to look like before the fonts have loaded? |
| 23:33 | <sayrer> | weinig_, no idea. I was using Firefox and saw a flash of unfontedness, and checked out Safari 4 to see how you guys handled it |
| 23:34 | <weinig_> | sayrer: ah |
