| 01:02 | <Hixie> | hey i passed r3000 earlier without noticing |
| 07:28 | <jgraham> | html5lib now with a hg repo |
| 07:29 | <MikeSmith> | jgraham: cool |
| 07:29 | <MikeSmith> | jgraham: google code provides hg as an option? |
| 07:36 | <jgraham> | MikeSmith: Yeah, it is experimental at the moment |
| 07:36 | <jgraham> | I haven't actually tried yet to se if it works :) |
| 07:38 | <MikeSmith> | jgraham: I notice on Google Code blog: |
| 07:38 | <MikeSmith> | "given that Google Code's infrastructure is built for HTTP-based services, we found that Mercurial had the best protocol and performance characteristics for HTTP support. For more information, see our analysis." |
| 07:38 | <MikeSmith> | that's an interesting statement |
| 07:38 | <MikeSmith> | I wonder what the git people would say about that |
| 07:43 | <jgraham> | MikeSmith: Maybe they would say "In a market with two rather similar, both popular, products they did careful analysis and concluded that the other product was a better fir for their needs. Bravo on such dilligence in assessing the best solution!". But I doubt it somehow |
| 07:43 | <MikeSmith> | jgraham: their http://code.google.com/p/support/wiki/DVCSAnalysis doc is worth reading |
| 07:45 | <MikeSmith> | I pointed W3C system team at it; maybe will help inform decisions about what DCVS to use at W3C |
| 07:50 | <jgraham> | MikeSmith: Yeah, I saw that. |
| 07:52 | <MikeSmith> | jgraham: but hey, no reason why that should let us choose to favor a "custom stateful protocol" rather than a stateless HTTP one |
| 07:53 | <MikeSmith> | I mean because, what evidence do we have to suggest that this stateless HTTP approach to things really works? |
| 07:53 | <MikeSmith> | we need to remain skeptical until we see some proof of that |
| 07:54 | <MikeSmith> | in the mean time, we can do like those guys did: Assume we know better, and invent more custom stateless stuff. |
| 08:32 | zcorpan | does not understand role="math" |
| 08:46 | <jgraham> | zcorpan: In what sense? |
| 08:49 | <jgraham> | The use case seems to be <img src="equation.gif" role="math" alt="\int_0_y x^{2} \d x" aria-describedBy="a"><p id="a">The integral between 0 and y of x squared dx</p> |
| 08:51 | <jgraham> | Of which only the last part is really silly because e.g. a LaTeX to HTML convertor could produce both equaton.gif and the alt text, and there is AT that understands some LaTeX |
| 08:51 | <jgraham> | Although it might not do much for non AT users and would probably not work in any reasonably complex LaTeX scenario with e.g. custom macros |
| 09:28 | <zcorpan_> | jgraham: will you post the --!> and --\s*> findings to the list? |
| 09:34 | <zcorpan_> | jgraham: apparently role="math" is appropriate for MathML, too |
| 09:35 | annevk2 | has the feeling aria-describedby is way too generic to succeed |
| 09:46 | <MikeSmith> | annevk2: please convey my appreciation to whomever it was that made the opera retro page happen |
| 09:46 | <MikeSmith> | that thing is great |
| 09:46 | <MikeSmith> | they just should have kept it there for longer |
| 09:47 | <MikeSmith> | but glad that I caught it when it was live |
| 09:47 | <annevk2> | it's still there for me |
| 09:48 | <MikeSmith> | hmm, it's showing the normal start page for me now |
| 09:48 | <annevk2> | if you tweak the URL you should get it back |
| 09:49 | <MikeSmith> | OK |
| 09:50 | <MikeSmith> | anyway, the Opera spirit is great |
| 09:50 | <takkaria> | I like Opera |
| 09:50 | <MikeSmith> | the humor, not taking itself deadly seriously all the time |
| 09:50 | <takkaria> | they offered me an internship |
| 09:51 | <MikeSmith> | takkaria: not surprised. Opera definitely has a knack for finding and recruiting smart people |
| 09:53 | <Philip`> | MikeSmith: Or they just offer internships to everybody :-) |
| 09:53 | <annevk2> | congrats takkaria |
| 09:54 | <MikeSmith> | Philip`: hmm, they never offered me an internship |
| 09:54 | <MikeSmith> | I feel left out |
| 09:56 | <zcorpan_> | MikeSmith: you can apply for one |
| 09:56 | <MikeSmith> | cool, I didn't know I still had time |
| 09:57 | <MikeSmith> | zcorpan_: how much does the gig pay? |
| 09:57 | <takkaria> | annevk2: cheers |
| 09:57 | <zcorpan_> | MikeSmith: no idea |
| 09:57 | <takkaria> | does anyone here work in the Linköping office? |
| 09:58 | <zcorpan_> | me and jgraham |
| 09:59 | <takkaria> | oh, great :) |
| 10:01 | <MikeSmith> | zcorpan_: do people in Linköping office still use that Emacs-based messaging thing instead of IRC? |
| 10:01 | <MikeSmith> | something-COM or COM-something |
| 10:02 | <zcorpan_> | MikeSmith: yes, LysKOM |
| 10:02 | <MikeSmith> | zcorpan_: do you use it? |
| 10:02 | <zcorpan_> | MikeSmith: yes i have to :) |
| 10:03 | <jgraham> | Yay LysKOM |
| 10:03 | <MikeSmith> | I guess I should instead ask, do you like using it? |
| 10:03 | <zcorpan_> | although i read it just once a week or so |
| 10:03 | <zcorpan_> | no i hate it |
| 10:03 | <jgraham> | Reinventing the wheel but making it triangular and spiky |
| 10:03 | <MikeSmith> | heh |
| 10:03 | <MikeSmith> | good description |
| 10:03 | <MikeSmith> | I don't know what they like about that thing |
| 10:04 | <MikeSmith> | I mean, I can see it being having some use |
| 10:04 | <MikeSmith> | but not as an IRC replacement |
| 10:04 | <zcorpan_> | it's something in between irc and email |
| 10:05 | <MikeSmith> | zcorpan_: right, but more of a worst-of-both-worlds combination of the two |
| 10:05 | <zcorpan_> | don't disagree |
| 10:06 | <jgraham> | The choice of clients seems to be a) a webbased client that breaks the back button and intermitttently gets confused about whether there are new messages or b) an emacs-based client |
| 10:07 | <MikeSmith> | jgraham: there's a curses-based client too, I seem to remember |
| 10:07 | <MikeSmith> | I think that's what I used, when I needed to |
| 10:07 | <jgraham> | Really? |
| 10:07 | <MikeSmith> | yeah, somewhere |
| 10:08 | <MikeSmith> | I think I found it less painful than Emacs |
| 10:09 | <MikeSmith> | http://www.lysator.liu.se/lyskom/klienter/ttyklient/ |
| 10:09 | <MikeSmith> | I guess |
| 10:11 | <MikeSmith> | http://packages.ubuntu.com/dapper/net/lyskom-tty-client |
| 10:13 | <MikeSmith> | is hsivonen site using downloadable fonts? |
| 10:13 | <MikeSmith> | 3MB and counting... |
| 10:14 | <MikeSmith> | 4MB+ |
| 10:16 | <MikeSmith> | hsivonen: anyway, you updated http://hsivonen.iki.fi/doctype/ today, I guess? |
| 10:17 | MikeSmith | sees <!DOCTYPE html> among the recommendations there now |
| 10:17 | <Philip`> | It sounds somewhat ironically like the Emacs client elicits more curses than the curses-based client |
| 10:21 | <MikeSmith> | heh |
| 10:21 | <MikeSmith> | actually, I see now it just says "TTY-based client" |
| 10:22 | <MikeSmith> | and package seems to depend only on libc6 package |
| 10:22 | <MikeSmith> | man, hsivonen got a truckload of downloadable fonts on his pages |
| 10:23 | <Philip`> | Not as many as on my page, I'm sure |
| 10:24 | <jgraham> | MikeSmith: It also says "With lyskom you can connect to your favourite LysKOM" |
| 10:24 | <MikeSmith> | well, your page has a specific demonstration purpose |
| 10:24 | <MikeSmith> | heh |
| 10:24 | <MikeSmith> | "your favourite LysKOM" |
| 10:24 | <MikeSmith> | that kind of limits things |
| 10:24 | <MikeSmith> | to zero |
| 10:24 | <MikeSmith> | zero choices there |
| 10:25 | <jgraham> | Indeed. It seems entirely unsutiable for my purposes |
| 10:25 | <Philip`> | "favourite" is relative |
| 10:25 | <Philip`> | like how you could have a favourite evil dictator |
| 10:25 | <MikeSmith> | "With lyskom you can connect to your least-despised LysKOM" |
| 10:26 | <MikeSmith> | "With lyskom you can connect to the LysKOM you have a need to use even though you hate it" |
| 10:27 | <MikeSmith> | so I'm wondering if the size/number of fonts at hsivonen site is atypical, or if a typical site that uses downloadable is going to be in similar range |
| 10:27 | <MikeSmith> | ~5.1MB |
| 10:28 | <zcorpan_> | depends on whether they use Philip`'s service or not |
| 10:28 | <MikeSmith> | or 6.4MB actually, on my Mac |
| 10:28 | Philip` | wonders how he could make his service more usable |
| 10:29 | <Philip`> | I suppose it'd be nice if the TTF+EOT thing wasn't such a horrid hack that results in IE making a 404 request for every font |
| 10:29 | <MikeSmith> | speaking of usable, Web Inspector UI is great for getting at this info |
| 10:29 | <MikeSmith> | download time/speed, I mean |
| 10:30 | <Philip`> | Also I suppose it'd be nice if you could ask for e.g. all ASCII letters, or all European letters, rather than having to type in exactly the characters you want |
| 10:30 | <Philip`> | MikeSmith: Better than Firebug? |
| 10:30 | <MikeSmith> | Philip`: I like it better at least |
| 10:31 | <Philip`> | Firebug handily alerted me to the issue that my pages were all taking 15 seconds to finish loading |
| 10:31 | <MikeSmith> | I really wish mozilla would just make Firebug a standard part of Firefox |
| 10:32 | <MikeSmith> | or at least of Minefield |
| 10:32 | <Philip`> | (since I had a (Fast)CGI script outputting some text/html with a Content-Length, and Apache was set to automatically compress text/html, but it doesn't change the Content-Length when compressing, so Squid was spending 15 seconds waiting for more content before timing out, I think) |
| 10:33 | jgraham | got a couple of crashes when he trie to use WI with Safari 4 |
| 10:34 | <Philip`> | MikeSmith: That sounds like needless bloat for a hundred million users who are never going to do any web development |
| 10:34 | <MikeSmith> | Philip`: yeah, I suppose so, for the FF case at least |
| 10:35 | <MikeSmith> | but seems it has some obvious practical use for bundling in Minefield |
| 10:35 | <Philip`> | MikeSmith: People who willingly install Minefield ought to know how to install an extension - it only takes a minute when you're first setting up your profile |
| 10:36 | <jgraham> | The big problem is when it is incompatible with minefield.latest |
| 10:36 | <MikeSmith> | right |
| 10:36 | <Philip`> | Then use some other extension that makes extensions compatible |
| 10:36 | <MikeSmith> | and I bet there are still people who don't install Firebug with Minefield anyway |
| 10:36 | <Philip`> | like https://addons.mozilla.org/en-US/firefox/addon/6543 |
| 10:36 | <jgraham> | Philip`: No, when it is actually incompatible |
| 10:38 | <Philip`> | jgraham: In that case bundling it with Minefield wouldn't help, because it would still be incompatible until somebody fixed it |
| 10:38 | <zcorpan_> | Philip`: you should make your service more discoverable using google, too |
| 10:38 | <MikeSmith> | Philip`: exactly what you said |
| 10:39 | <Philip`> | zcorpan_: But that might result in people using it, which would be bad because it's inefficient and buggy and ugly :-) |
| 10:39 | <MikeSmith> | a change that breaks Firebug is a change that should be backed out before the nightly is built |
| 10:39 | <jgraham> | Philip`: It would help because it wwould have to be treated at the same priority level as other features |
| 10:40 | <Philip`> | jgraham: It could be treated at the same priority level as other features without being bundled with Minefield |
| 10:41 | <Philip`> | by appropriate changes to development policy and to the automatic tests |
| 10:41 | <jgraham> | Philip`: Sure. But it is less likely to actually happen |
| 10:41 | <zcorpan_> | Philip`: maybe you could use spellchecking dictionaries to determine which characters to include for different languages |
| 10:41 | <jgraham> | zcorpan_: Punctuation |
| 10:41 | <Philip`> | jgraham: But more likely to actually happen than bundling with Minefield |
| 10:42 | <Philip`> | since it's a necessary step before bundling with Minefield, and therefore a subset of the challenge |
| 10:43 | <Philip`> | zcorpan_: Or look through the textual content of web pages that use different languages, to see what characters they use |
| 10:43 | <Philip`> | (which would be better at finding punctuation and digits and ASCII and copyright symbols, I guess) |
| 10:45 | <Philip`> | But it'd probably be easier to just go by Unicode blocks |
| 10:47 | <zcorpan_> | i guess the copyright symbol should be present for all languages |
| 10:47 | <Philip`> | Not if you only want to use the font for your logo, though |
| 10:47 | <zcorpan_> | right |
| 10:48 | <Philip`> | Perhaps what we should do is invent a new dynamic font format, so the web browser can download glyphs automatically on-demand without having to download irrelevant junk |
| 10:50 | <Philip`> | (EOT seems to have a primitive form of that idea, by copying various font properties to a header at the start of the file, so the browser can simply download the first chunk and decide whether it can use this font (because it covers the right character range, or has the right font style, etc) instead of having to download the entire file and then search through it for the same information in arbitrary unpredictable locations) |
| 10:59 | <Philip`> | Hmm, it looks like people have actually used my font thing |
| 10:59 | <Philip`> | since there's ~500 fonts in the output cache |
| 11:00 | <zcorpan_> | Philip`: could be people just trying it without actually using the fonts |
| 11:00 | <zcorpan_> | Philip`: btw the output page doesn't have a <title> |
| 11:00 | <zcorpan_> | last i checked |
| 11:03 | <Philip`> | zcorpan_: Oops |
| 11:05 | Philip` | fixes |
| 13:04 | <annevk2> | MikeSmith: http://www.opera.com/?flashed=0 lets you revisit the home page |
| 13:04 | annevk2 | wonders if home pages are archived somewhere |
| 13:24 | <hendry> | Hixie: noticed a typo in html5 index, %s/initalize/initialize/ |
| 13:39 | <zcorpan_> | <div class="float-left padding-top-5 padding-bottom-9 width-385"> -- http://www.thetimes.co.uk |
| 13:43 | <Philip`> | zcorpan_: They're not the only ones - e.g. www.grownups.co.nz/profile/viewPublicProfile/id/8209 has a class="border-left-blue border-right-blue padding-left-10 padding-right-10 padding-bottom-5" |
| 13:43 | <Philip`> | www.paintingmax.com/shop-style-ashcan-school-c-3_304.html?sort=2d - class="padding-10px" |
| 13:44 | <Philip`> | surfingshot.com/index.php?main_page=product_info&products_id=4432 - class="padding-2px" |
| 13:44 | <Philip`> | etc |
| 13:44 | Philip` | wonders if these sites picked up that trick from a common source |
| 13:50 | <Philip`> | hsivonen: http://hsivonen.iki.fi/doctype/ says "emumerate" |
| 13:51 | <Philip`> | hsivonen: "Testing image alignment with Internet Explorer is inadequate however be sure to test in IE8, too." is confusing to read, and maybe could do with some punctuation around the 'however' |
| 13:58 | <Philip`> | hsivonen: "To put it another way, if you have #include "foo.h", you should not bind any black magic to the name foo.h, because it should be permissible to paste the contents of foo.h inline or copy the contents of foo.h to bar.h and say #include "bar.h"." |
| 13:59 | <Philip`> | Modern compilers do do black magic when you have #include "foo.h", e.g. GCC will look for a precompiled header named foo.gch |
| 14:00 | <Philip`> | and if foo.h contains "#pragma once" then MSVC will treat #include "foo.h"; #include "foo.h" differently to how it treats #include "foo.h"; #include "differently-named-copy-of-foo.h" |
| 14:01 | <Philip`> | I suppose #pragma once is generally considered a bad idea, but precompiled headers a good idea, and they're definitely black magic |
| 14:01 | <Philip`> | (but they're magic which only affects performance, not functionality, which seems like a significant difference) |
| 14:11 | zcorpan_ | wonders why testing image alignment in IE8 is inadequate |
| 14:45 | <krijnh> | Philip`: sup? :) |
| 16:16 | <annevk2> | hsivonen, there's <table><tr><td><p><\/tbody> <\/table> in a document.write in Acid3 that assumes the space is still there after parsing |
| 16:16 | <annevk2> | hsivonen, Opera initially failed that |
| 16:16 | <hsivonen> | grrr! |
| 16:16 | <annevk2> | hsivonen, not a 100% sure it affects what you're discussing but I thought I'd bring it up |
| 16:17 | <hsivonen> | annevk2: what's "there" |
| 16:18 | <annevk2> | hsivonen, bottom of the page has the document.write |
| 16:18 | <hsivonen> | inside table? |
| 16:18 | <annevk2> | hsivonen, it writes a <table> including a space |
| 16:20 | gsnedders | has his internet get really slow |
| 16:20 | <hsivonen> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/98 is the same for me in Gecko/HTML5, WebKit and Opera |
| 16:22 | <annevk2> | since Opera 10, yes |
| 16:24 | <annevk2> | we used to ignore whitespace at certain places in tables during parsing |
| 16:36 | <hendry> | is there an EU mirror of whatwg.org/html5? :) |
| 16:38 | <Philip`> | hendry: What's wrong with America? It's only a hundred milliseconds away :-) |
| 16:41 | hendry | wonders how http://svn.whatwg.org/webapps/index is split up to multipage |
| 16:42 | <MikeSmith> | hendry: Philip` does it manually every time Hixie commits |
| 16:44 | <hendry> | hah |
| 16:45 | <MikeSmith> | hendry: |
| 16:45 | <MikeSmith> | http://code.google.com/p/html5/source/browse/trunk/spec-splitter/spec-splitter.py |
| 16:46 | <MikeSmith> | http://html5.googlecode.com/svn/trunk/spec-splitter/spec-splitter.py |
| 16:48 | <Philip`> | hendry: That spec-splitter script runs on my server, and Hixie executes it and then downloads the output |
| 16:50 | <jgraham> | Don't believe Philip`; that script is just for show. He actually has a highly trained team of bees that read the spec and output the html for the split spec as a pattern of filled an unfilled cells in their hive |
| 16:50 | <Philip`> | I love bees |
| 16:55 | <hendry> | Philip`: don't laugh now, did you try xsl? |
| 16:56 | jgraham | wonders wwhen he should laugh |
| 16:59 | <jgraham> | How can xsl possibly compete with the massive concurrency of a whole swarm of bees |
| 16:59 | <jgraham> | ? |
| 17:04 | <hendry> | i ran xsltproc on html5's index today to extract pre[@class='idl'] and I was impressed how fast it was. |
| 17:05 | <Philip`> | hendry: The difficulty with XSL is that I don't understand it |
| 17:06 | <Philip`> | particularly for complex things like splitting a document in multiple parts and gluing the right headers and inter-page navigation and fixing hashed links and stuff |
| 17:06 | <Philip`> | (and particularly since I've never cared to try learning XSL) |
| 17:07 | <Philip`> | whereas I had already learned Python and knew how to use it |
| 17:08 | <gsnedders> | Philip`: But parsing the spec is slow in Python. |
| 17:08 | <Philip`> | gsnedders: No it's not, since I use lxml's parser |
| 17:08 | <gsnedders> | Cheat. |
| 17:08 | <Philip`> | It's all your fault for not making html5lib fast enough |
| 17:09 | <gsnedders> | I'm not the one who was optimizing it. |
| 17:09 | <jgraham> | Philip`: Isn't that your fault? I thought thte deal was that I worked on making it slower, you worked on making it faster, gsnedders hecked from the sidelines? |
| 17:09 | <Philip`> | gsnedders: That's why it's your fault |
| 17:10 | <gsnedders> | jgraham: s/hecked/hacked/? |
| 17:10 | <Philip`> | heckled |
| 17:10 | <jgraham> | heckled |
| 17:10 | <gsnedders> | Ah, I wasn't sure which |
| 17:10 | <gsnedders> | I guessed it was one of them |
| 17:10 | <jgraham> | Raeding what you write on irc is way overrated |
| 17:10 | gsnedders | turns up the volume of Life in Cartoon Motion |
| 17:11 | Philip` | feels like a proper researcher now since he's reading archeologically ancient papers - some are from as far back as 1999, and published as PostScript rather than PDF |
| 17:12 | gsnedders | is listening to Lollipop by Mika from Life In Cartoon Motion |
| 17:12 | gsnedders | is blatantly cool |
| 17:12 | <MikeSmith> | hendry: I've been planning to write an XSLT-based splitter for the H:TML draft |
| 17:12 | <jgraham> | Philip`: Isn't ps irritating though? |
| 17:13 | <Philip`> | jgraham: Yes |
| 17:13 | gsnedders | goes back to trying to understand solving first order linear differential equations |
| 17:13 | <Philip`> | jgraham: mainly since I can't read them in KPDF, and have to use something weird and crazy like KGhostView which doesn't handle page-breaks nicely |
| 17:14 | <hendry> | MikeSmith: get anywhere with that? ;) |
| 17:15 | <hendry> | Philip`: i'm trying now to minimize your alpha canvas test for inclusion to the MCTMB. http://static.webvm.net/canvas/2d.drawImage.alpha.html |
| 17:16 | <MikeSmith> | hendry: not yet, but I suppose I should turn my attention to it now. I would need to eventually anyway |
| 17:16 | <hendry> | Philip`: going to give up for now as I need to catch a train. Need to understand your assert thingie. Just wanted to update http://www.w3.org/2008/06/mobile-test/canvas.js |
| 17:17 | <Philip`> | hendry: The _assertPixelApprox is just requiring that the output is green |
| 17:18 | <MikeSmith> | only troublesome part about splitting with XSLT is how to figure out which fragment refs are going to end up being intra-file ones, and dealing with those |
| 17:18 | <Philip`> | (with some tolerance, because implementations might not be precise) |
| 17:18 | <MikeSmith> | might be a good case for using XSL kyes |
| 17:18 | <Philip`> | (and only checking one pixel of the image, because it's too slow and pointless to check every pixel) |
| 17:19 | <Philip`> | hendry: If your test output is checked by humans then you shouldn't need to do anything fancy like that - just tell people the output should look green |
| 17:20 | <hendry> | Philip`: yes, that's what I am trying to do. That's haot the MCTMB test already works: http://www.w3.org/2008/06/mobile-test |
| 17:42 | <gsnedders> | Oh, wow. This isn't as hard as I thought. |
| 17:43 | gsnedders | is having vague understanding creep into his head |
| 17:48 | Philip` | wonders why none of his PS viewers have a search function |
| 17:49 | <Philip`> | (I'm guessing it's just because PS is too stupid a document format for such things) |
| 20:08 | <cryzed> | Hey |
| 20:08 | <cryzed> | just wanted to drop by |
| 20:08 | <cryzed> | and tell your that's something wrong with the Google |
| 20:09 | <cryzed> | mercurial |
| 20:09 | <cryzed> | repository |
| 20:09 | <cryzed> | If you haven't already noticed |
| 20:09 | <jgraham> | Which |
| 20:09 | <jgraham> | one |
| 20:09 | <jgraham> | ? |
| 20:09 | <cryzed> | http://code.google.com/p/html5lib/ |
| 20:09 | <jgraham> | Ah |
| 20:09 | <cryzed> | You can't even browse the source |
| 20:09 | <cryzed> | via the web interface |
| 20:09 | jgraham | should investigate |
| 20:09 | <cryzed> | http://code.google.com/p/html5lib/source/browse/ -> Gives you an error |
| 20:09 | <cryzed> | "Error retrieving directory contents." |
| 20:09 | Philip` | wonders if it was wise to switch to such an experimental system :-) |
| 20:10 | <jgraham> | Philip`: Well we have other source trres around, right? |
| 20:10 | <jgraham> | :) |
| 20:11 | <Philip`> | jgraham: Do we? |
| 20:12 | <jgraham> | Philip`: Sure. Just not with revision history and stuff |
| 20:12 | <jgraham> | Which is useful but sort-of non essential |
| 20:14 | <jgraham> | It might be that I need to do something |
| 20:14 | <jgraham> | I am investigating now |
| 20:15 | <gsnedders> | jgraham: Are we meant to all be moving to hg now for html5lib? |
| 20:15 | <gsnedders> | http://html5lib.googlecode.com/hg/ looks sorta wrong |
| 20:16 | <jgraham> | gsnedders: That is sort of the plan |
| 20:17 | <jgraham> | cryzed, Philip`, gsnedders The svn repo is still avaliable at https://html5lib.googlecode.com/svn/trunk/ |
| 20:18 | <jgraham> | It seems like I need to import the svn repositoy into hg |
| 20:19 | <cryzed> | thanks |
| 20:19 | jgraham | wishes OSX had sane package mangement, decides to move to a linux computer |
| 20:21 | <cryzed> | jgraham, high five! |
| 20:21 | <cryzed> | Ubuntu <3 |
| 20:21 | <cryzed> | Well not necessarily Ubuntu, any Linux will do :D |
| 20:21 | <Philip`> | Except Slackware |
| 20:21 | <cryzed> | Except Gentoo |
| 20:22 | <Philip`> | Gentoo has sane package management |
| 20:22 | <cryzed> | Gentoo's compile-time is just too fucking long |
| 20:22 | <cryzed> | Why would I want to waste hours compiling X11 and GNOME? |
| 20:23 | <Philip`> | It's not wasting time - you can multitask |
| 20:23 | <cryzed> | imho it is |
| 20:23 | <cryzed> | You can simply use pre-compiled binaries |
| 20:23 | <Philip`> | You can write "emerge -u xorg-server" and then it'll be finished long before you next reboot |
| 20:24 | <cryzed> | yeah, and what do you do without a GUI? |
| 20:24 | <cryzed> | Editing text-files? |
| 20:24 | <cryzed> | Lynx? |
| 20:24 | <cryzed> | :D |
| 20:24 | <Philip`> | You install from the precompiled binaries on the installation CD |
| 20:24 | <Philip`> | and then once you've got a running system you can keep it up-to-date in the background |
| 20:25 | <cryzed> | there are precompiled xorg binaries? |
| 20:25 | <cryzed> | (and gnome?) |
| 20:25 | <Philip`> | There's a whole Live CD with GUI that you can install from |
| 20:25 | <cryzed> | Well, yes the LiveCD hast got a GUI |
| 20:26 | <cryzed> | But the installed operating system hasn't |
| 20:26 | <cryzed> | right? |
| 20:26 | <gsnedders> | It depends if you install it. |
| 20:26 | <gsnedders> | There are binaries available for a lot of packages. |
| 20:26 | <cryzed> | gsnedders, So where exactly is the advantage |
| 20:26 | <cryzed> | over Arch and Ubuntu? |
| 20:26 | <cryzed> | You can compile your programs |
| 20:26 | <cryzed> | ok cool |
| 20:27 | <cryzed> | 1-2 seconds execution time boost |
| 20:27 | <cryzed> | I guess |
| 20:27 | <Philip`> | You can get continually up-to-date software, rather than waiting six months for the next big release |
| 20:27 | <Philip`> | You can customise the build options and compile-time features of all the software |
| 20:27 | <cryzed> | Ubuntu's updating regularly aswell |
| 20:27 | <gsnedders> | Better way of speeding up start up time: switch to FreeBSD, which has always been more responsive for me than any Linux distro. |
| 20:28 | <cryzed> | gsnedders, Ah I'm curious |
| 20:28 | <cryzed> | I never used FreeBSD |
| 20:28 | <cryzed> | FreeBSD is basically |
| 20:28 | <cryzed> | another kernel |
| 20:28 | <cryzed> | right? |
| 20:28 | <gsnedders> | No, it's a complete OS |
| 20:28 | <cryzed> | Does it use GNU? |
| 20:28 | <gsnedders> | No |
| 20:28 | <cryzed> | So it isn't compatible |
| 20:28 | <cryzed> | with Linux binaries? |
| 20:28 | <gsnedders> | It has a binary compatibility module |
| 20:29 | <gsnedders> | (kernel module, that is) |
| 20:29 | <cryzed> | but still |
| 20:29 | <cryzed> | I don't know |
| 20:29 | <cryzed> | I think I prefer Ubuntu, seems easier and I'm by nature a lazy person |
| 20:30 | <cryzed> | Not stupid, I probably could switch if I wanted, but I don't |
| 20:30 | <cryzed> | I'm rather content with Ubuntu and the boot-time |
| 20:30 | <cryzed> | (about 18 seconds) |
| 20:30 | gsnedders | has plenty of problems with most DEs' usability |
| 20:30 | <gsnedders> | (where most is defined to be everyone I've tried) |
| 20:30 | <gsnedders> | (this is X DEs) |
| 20:31 | <cryzed> | You are using X DE? |
| 20:31 | <gsnedders> | ? |
| 20:31 | <cryzed> | nevermind |
| 20:31 | <cryzed> | I didn't understand what you said |
| 20:31 | <gsnedders> | an X desktop enviroment |
| 20:31 | <gsnedders> | a desktop environment for an X window server |
| 20:32 | <cryzed> | ah okay |
| 20:32 | <cryzed> | Well I like GNOME |
| 20:32 | <cryzed> | I hate KDE |
| 20:32 | <cryzed> | XFCE is not my thingy |
| 20:32 | <cryzed> | never tried any *box |
| 20:34 | gsnedders | sticks to OS X because it has a better desktop environment and still has a proper POSIX command line |
| 20:36 | jgraham | wonders what problems gsnedders has |
| 20:36 | gsnedders | has a lot of problems |
| 20:37 | <jgraham> | with X desktop environments in particula |
| 20:37 | <jgraham> | r |
| 20:37 | <gsnedders> | Then tend to just suck once you get to the small details |
| 20:38 | <gsnedders> | Now, unless this helps me with maths, this is about as productive as a ball. |
| 20:39 | <Philip`> | Why don't you think this is productive? It's an opportunity to end the OS/desktop debate once and for all, by coming to a decision on which is best - you can't just give up now |
| 20:44 | <gsnedders> | Philip`: I don't have an exam on this. |
| 20:45 | gsnedders | needs to get a buttonhole |
| 20:45 | <gsnedders> | (as in, the floral thing, not the actual hole) |
| 20:45 | <Philip`> | gsnedders: Sure, but think of how much humanity will benefit once we've decided which OS is best and so nobody else will have to spend time debating it |
| 20:46 | <gsnedders> | "There's many battles lost/But tell me who has won" |
| 20:51 | <jgraham> | Plenty of OS X details suck too |
| 20:51 | <jgraham> | :) |
| 20:51 | <cryzed> | Ehrm |
| 20:51 | <cryzed> | Guys |
| 20:51 | <cryzed> | Linux is best |
| 20:51 | cryzed | is running |
| 20:52 | <cryzed> | btw, thanks for the great library |
| 20:52 | <cryzed> | using it daily |
| 20:52 | <cryzed> | to do my webscraping with python |
| 20:57 | gsnedders | uses what he thinks is best for what he needs to do |
| 20:57 | <gsnedders> | Which tends to be for any computer I have a GUI on OS X |
| 20:58 | Philip` | does the same, and ends up using Linux, and the difference is that he is right and gsnedders is wrong |
| 21:01 | jgraham | tendds to use anything that isn't Windows as he finds that too annoying and everything else merely differntly deficient |
| 21:02 | gsnedders | is getting to the point where he can actually solve first order linear differential equations without looking at how to do it and get the answer corret |
| 21:02 | <gsnedders> | *correct |
| 21:03 | <jgraham> | Examples of things that are broken on OS X: The lack of control over window positioning. The applications folder. The broken virtual desktops. The finder. |
| 21:04 | <jgraham> | gsnedders: good :) |
| 21:05 | <Philip`> | jgraham: Also, it looks too pretty |
| 21:05 | <gsnedders> | Virtual desktops have been far less broken since 10.5.3, though |
| 21:05 | <gsnedders> | Like, they are actually bearable and useful now |
| 21:05 | <Philip`> | I like to be reminded that I'm using a deterministic machine and I have dominion over it |
| 21:06 | <jgraham> | gsnedders: Not with Lightroom |
| 21:06 | <gsnedders> | How so? |
| 21:06 | <jgraham> | (this may not be Apple's fault. I don't care) |
| 21:07 | <jgraham> | gsnedders: Click LR icon in the dock. Dock dissapears. Desktop containing LR is not focused |
| 21:08 | <jgraham> | (assuming i is open and on a different virtual desktop) |
| 21:08 | gsnedders | assumes that LR creates its own windows and doesn't use any standard API for it |
| 21:09 | <gsnedders> | But yeah, that sounds b0rked |
| 21:17 | Philip` | got confused when switching from desktop A to desktop B resulted in all the terminal windows on desktop A jumping into the foreground, and then gave up |
| 23:48 | <mathuin> | What is the best way to normalize all the tags in a BeautifulSoup tree generated by html5lib? Some files I use have tags like 'dc:Title', others 'dc:title' so I'd like to force all tags to lower case. Do I use a sanitizer or something else? |