00:00:08 -!- sebbu has joined.
00:13:15 <AnMaster> ais523, link?
00:15:14 <ais523> AnMaster: I saw it on Slashdot
00:15:17 <ais523> it's probably still there
00:15:46 <ais523> the links are http://marketshare.hitslink.com/status.aspx and http://gs.statcounter.com/#browser-na-daily-20080701-20090707, apparently, but I didn't follow them
00:15:58 <ais523> people don't actually follow links at Slashdot, they just read the summary
00:16:24 <AnMaster> last I looked slashdot was large
00:16:57 <pikhq> Slashdot is fucking huge.
00:19:11 <AnMaster> so where on slashdot ais523
00:19:19 <ais523> AnMaster: main page
00:19:25 <ais523> probably near the top
00:19:51 <ais523> third entry atm
00:23:09 -!- lambdabot has joined.
00:24:15 -!- BeholdMyGlory has quit (Remote closed the connection).
00:26:32 -!- BeholdMyGlory has joined.
00:26:45 -!- BeholdMyGlory has quit (Remote closed the connection).
00:27:06 <oerjan> o_<
00:27:09 <AnMaster> I suspect I know the cause
00:27:29 <AnMaster> new major firefox version release. Thus some clients counted more than once, once for old user agent, once for new
00:27:33 <AnMaster> ais523, plausible?
00:27:46 <ais523> ooh, could be plausible
00:27:57 <ais523> except that it was a drop in IE, rather than a rise in Firefox
00:27:57 <oerjan> > fun.cycle$"lambda! "::Expr
00:27:59 <lambdabot>   lambda! lambda! lambda! lambda! lambda! lambda! lambda! lambda! lambda! lam...
00:28:00 <thutubot>   lambda! lambda! lambda! lambda! lambda! lambda! lambda! lambda! lambda! lam...
00:28:03 <ais523> the drop was filled by Firefox, Safari and Chrome
00:28:06 <AnMaster> ais523, ah ok
00:28:10 <oerjan> what the?
00:28:13 <AnMaster> so not just one
00:28:49 <oerjan>    test
00:29:12 <oerjan> thutubot: > 2+2
00:29:21 <oerjan> lambdabot: > 2+2
00:29:29 <oerjan> grmbl
00:29:35 <oerjan> lambdabot: @run 2+2
00:29:36 <lambdabot>   4
00:29:37 <thutubot>   4
00:29:43 <oerjan> bah
00:29:57 <oerjan> hm that actually proves nothing
00:30:16 <oerjan>    4
00:30:35 <AnMaster> um
00:30:46 <AnMaster> ais523 might be able to explain it
00:30:57 <AnMaster> iirc thutubot is his bit
00:30:58 <AnMaster> bot*
00:31:00 <oerjan> oh wait
00:31:09 <AnMaster>   4
00:31:11 <oerjan> @run 2+2
00:31:12 <lambdabot>   4
00:31:12 <thutubot>   4
00:31:29 <oerjan> @vixen who are you?
00:31:29 <lambdabot> what do you mean, i'm me!
00:31:30 <thutubot> what do you mean, i'm me!
00:31:35 <AnMaster> err
00:31:43 <oerjan> @vixen how do you feel?
00:31:43 <lambdabot> however you want
00:31:44 <thutubot> however you want
00:31:47 <AnMaster> ais523, care to explain what the hell is going on
00:32:03 <oerjan> @vixen Microoptimisation of gnomes
00:32:03 <lambdabot> wanna hear a story?
00:32:03 <thutubot> wanna hear a story?
00:32:22 <oerjan> hm lessee
00:32:24 <ais523> AnMaster: someone's /msging thutubot
00:32:35 <thutubot> like this
00:32:35 <AnMaster> ais523, that quickly?
00:32:36 <oerjan> @dice 1d3
00:32:36 <lambdabot> 1d3 => 3
00:32:37 <thutubot> 1d3 => 3
00:32:44 <AnMaster> ais523, how strange
00:32:54 <AnMaster> why would someone set up a lambdabot -> thutubot relay?
00:33:01 <ais523> because it's funny?
00:33:01 <AnMaster> ais523, can you tell us who?
00:33:14 <ais523> and not easily, thutubot isn't saving debug data anywhere
00:33:18 <ais523> AFAIR
00:33:19 <AnMaster> ais523, not particularly
00:33:28 <AnMaster> ais523, stdout?
00:33:46 <AnMaster> also how did you produce the "like this"?
00:33:59 <AnMaster> as in
00:34:01 <AnMaster> what command
00:34:06 <AnMaster> and prefix
00:34:33 <AnMaster> or just using +ul?
00:34:38 <AnMaster> +ul (test)S
00:34:38 <thutubot> test
00:34:45 <thutubot> test
00:34:45 <AnMaster> no
00:34:47 <AnMaster> wait yes
00:34:49 <AnMaster> heh
00:35:03 <ais523> ok, that is weird
00:35:06 <AnMaster> ais523, why not return it to sender?
00:35:12 <AnMaster> /msg thutubot +ul (test)S
00:35:18 <AnMaster> returns it to channel
00:35:20 <AnMaster> not sender
00:35:22 <ais523> thutubot is printing everything sent to it, and people haven't been /msging it until I suggested it
00:35:45 <AnMaster>  @dice 1d3
00:35:46 <AnMaster> @dice 1d3
00:35:47 <lambdabot> 1d3 => 2
00:35:47 <thutubot> 1d3 => 2
00:35:52 <AnMaster> ais523, what happened just there
00:36:25 <ais523> ok, that is weird; thutubot didn't see itself sending that
00:36:31 <ais523> +ul test
00:36:47 <ais523> oh, but it isn't seeing its own sends anyway
00:36:54 <ais523> ahaha, I've figured it out
00:37:00 <ais523> it's thutubot's +haskell command
00:37:04 <ais523> +haskell 2+2
00:37:15 <ais523> it's meant to send the message to lambdabot, then relay the response
00:37:18 <ais523> it's broken, though
00:37:26 <AnMaster> ais523, ouch
00:37:27 <oerjan> ah!
00:37:34 <ais523> in two ways: a) it isn't sending on to lambdabot, and b) it's repeating everything lambdabot says
00:37:36 <AnMaster> broken both ways
00:45:53 -!- ais523 has quit (Remote closed the connection).
00:50:10 -!- immibis has joined.
00:56:16 -!- FireFly has quit ("Later").
01:05:07 -!- darthnuri has quit (No route to host).
01:05:44 -!- inurinternet has quit (No route to host).
01:17:47 -!- inurinternet has joined.
01:17:53 -!- darthnuri has joined.
01:37:09 -!- darthnuri has quit (Connection timed out).
01:37:54 -!- inurinternet has quit (Connection timed out).
01:38:04 -!- immibis has quit ("ASCII a stupid question, get a stupid ANSI!").
01:59:07 -!- inurinternet has joined.
02:04:42 -!- darthnuri has joined.
02:09:38 -!- oerjan has quit ("Good night").
02:26:50 -!- sebbu2 has joined.
02:27:15 -!- oklodol has joined.
02:39:43 -!- sebbu3 has joined.
02:43:04 -!- sebbu has quit (Read error: 110 (Connection timed out)).
02:43:32 -!- sebbu has joined.
02:44:10 -!- oklodok has quit (Success).
02:48:27 -!- sebbu3 has quit (Read error: 60 (Operation timed out)).
02:57:13 -!- sebbu2 has quit (Read error: 110 (Connection timed out)).
04:29:14 -!- lifthrasiir has changed nick to arachneng.
04:29:26 -!- arachneng has changed nick to lifthrasiir.
04:57:18 -!- coppro has quit (Remote closed the connection).
05:03:00 -!- coppro has joined.
06:04:10 -!- Associat0r has joined.
06:15:54 <Deewiant> http://googleblog.blogspot.com/2009/07/introducing-google-chrome-os.html
06:15:56 <Deewiant> :-S
06:18:29 -!- coppro has quit (Remote closed the connection).
06:21:24 -!- coppro has joined.
06:30:54 -!- coppro has quit ("The only thing I know is that I know nothing").
06:32:09 -!- coppro has joined.
06:53:22 <chuck> Deewiant: just another crucial step in google's plot to take over the world.. muahaha.
07:02:24 -!- GuestShadowSkunk has joined.
07:15:08 -!- Slereah has quit (Read error: 110 (Connection timed out)).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:45:23 -!- chuck has left (?).
09:16:10 -!- FireFly has joined.
09:19:20 -!- Associ8or has joined.
09:20:02 <oklodol> fizzie: Five in GMP, I mean. It is rather unlikely there would be no others. <<< you can get an infinite descent of asymptotically better multiplication algos by extending the divide-and-conquer method of karatsuba; but i don't know how fft is done for numbers, if it's O(n lg n) it's beyond that descent
09:20:16 <oklodol> i thought fft only made sense for polys
09:21:04 <fizzie> http://numbers.computation.free.fr/Constants/Algorithms/fft.html  has a shortish explanation.
09:22:58 <fizzie> And actually the GMP manual explains what they do, too.
09:29:45 <oklodol> estoppel: pikhq: problem with digits is you have to pick a base, and that sucks. <<< you can use other kinds of series, like continued fractions; but really there's no need to have an explicit representation until you print
09:29:59 <oklodol> oerjan: estoppel: oh. in that case use continued fractions. no base involved. <<< right, i was already here
09:35:41 -!- M0ny has joined.
09:36:26 -!- Pthing has joined.
09:36:55 -!- Associat0r has quit (Connection timed out).
09:37:26 -!- coppro has quit (Read error: 110 (Connection timed out)).
09:44:04 <oklodol> fizzie: what year is that idea from?
09:44:37 <oklodol> i thought there was no known n lg n for multiplication, even though i knew fft does that for polynomials
09:45:52 <fizzie> The Schönhage and Strassen paper referred to in the GMP manual are from 1971, but I haven't actually read it, so I don't know what they promise w.r.t. multiplication.
09:47:13 <oklodol> hmm
09:47:34 <oklodol> okay right, seems you can reduce the multiplication to multiplication of polynomials
09:48:02 <fizzie> s/are/is/ there.
09:49:37 <oklodol> oh, that was the problem, i thought i saw something weird there
09:50:17 <oklodol> "and a final rearrangement on the coefficients of R(z) permits to obtain the product XY" what rearrangement?
09:51:19 -!- MigoMipo has joined.
10:02:05 <fizzie> I would peek at the 1971 paper, but based on the title -- "Schnelle Multiplikation grosser Zahlen" -- it is probably in German. And that other page refers to a paper called "Multiplication en le Lisp" in French.
10:02:29 <fizzie> I'm sure there's a less overviewy explanation in the tubes, though.
10:02:51 <AnMaster> <Deewiant> http://googleblog.blogspot.com/2009/07/introducing-google-chrome-os.html <-- Is it based on linux or from scratch?
10:02:53 <oklodol> i think i already decrypted the page
10:02:58 <oklodol> from scratch
10:03:07 <AnMaster> ah wait
10:03:11 <AnMaster> it says linux there
10:03:13 <oklodol> at least that's the impression i got
10:03:16 <oklodol> alright
10:03:26 <AnMaster> "The software architecture is simple — Google Chrome running within a new windowing system on top of a Linux kernel."
10:04:22 <AnMaster> So basically you can't use it where there is no internet connection I guess
10:04:27 <oklodol> well, i don't see the relevance of that info, so i guess i wouldn't remember the fact either.
10:06:46 <oklodol> AnMaster: it's made for "people who live on the net"
10:07:15 <AnMaster> oklodol, yes right, but what if you are traveling and there happen to be no public wlan where you are or such.
10:07:24 <AnMaster> and no mobile phone connection either
10:08:27 <oklodol> you'd think there's some offline happenings too. but i just took a glance at the public info thingie.
10:08:29 <fizzie> Then you should be getting out of such a dark place.
10:08:36 <fizzie> With rapidity.
10:08:42 <oklodol> also what fizzie said
10:08:57 <AnMaster> I was travelling by train from Kiruna (in north Sweden) down to the south parts of Sweden a few weeks ago. Except around the towns the mobile phone mostly reported no connection or extremely bad one.
10:09:06 <AnMaster> and I have Telia
10:09:08 <oklodol> you can't swim if internet don't stream
10:09:12 <AnMaster> which is known for good coverage.
10:09:50 <oklodol> statisticians say 97% of finland is covered by connective flow.
10:09:53 <fizzie> I don't remember when I've last been outside GPRS coverage; of course online activities at "a kilobyte every now and then" speeds might not be so pleasant.
10:10:02 <AnMaster> fizzie, that was outside GSM even!
10:11:01 <AnMaster> and since my phone tends to switch over to some other carrier and say "emergency calls only" if Telia isn't reachable but even that didn't happen, I assume outside all coverage.
10:13:08 <AnMaster> most of the distance between Kiruna and Boden there was no coverage at all
10:13:10 <fizzie> The page I found about the 97 % GSM coverage is from 2001, they might've improved that a tiny bit. Anyway, if you don't get the internets, it's your own fault for going to such a silly place.
10:13:20 <oklodol> umm
10:13:27 <oklodol> it actually *is* 97?
10:13:41 <fizzie> oklodol: It was, according to http://www.eduskunta.fi/faktatmp/utatmp/akxtmp/kk_945_2001_p.shtml
10:13:50 <oklodol> i basically just let the flow tell the coverage to me.
10:13:51 <AnMaster> fizzie, oh in Finland?
10:13:52 <fizzie> Well, for Sonera's GSM network.
10:13:55 <AnMaster> Not worldwide
10:13:59 <oklodol> i just felt it
10:14:06 <fizzie> Yes, I have no idea about other places.
10:14:39 <oklodol> i wonder if i can feel statistics correct nowadays.
10:14:42 <fizzie> Still, I'm sure the thing will have at least some offline functionality.
10:14:55 <AnMaster> in south Sweden I tend to get at least one "dot" on the connection quality indicator thingy on the phone
10:15:01 <AnMaster> everywhere
10:15:27 <fizzie> I guess there is probably still some out-of-GSM-coverage areas up there in Lapland.
10:15:38 <AnMaster> fizzie, quite a lot even
10:16:34 <oklodol> 3% of finland is not that much
10:16:48 <fizzie> Well, I don't know about "lot". They do claim that 97 % is geographically speaking, not "amount of people", so...
10:16:49 <AnMaster> In Sweden I meant
10:17:22 <fizzie> Well, start building base stations then!
10:18:34 <AnMaster> fizzie, right, probably will cover one house each at most... Up there you can go on for miles without seeing a single house, or even a road crossing the railway.
10:19:21 <fizzie> If we've managed to cover 97 % (at least) of our surface area, you should too.
10:19:28 <AnMaster> possibly worth it along the railroad/railway (what is the diff? Is one US and one UK?), but apart from that it wouldn't be economically viable I bet.
10:19:58 <oklodol> it's not about viability
10:20:03 <oklodol> it's about ability
10:20:29 <AnMaster> oklodol, who pays?
10:20:42 <oklodol> taxpayers?
10:21:13 <AnMaster> the gov owns Sonera?
10:21:34 <oklodol> i suggest taxpayers pay straight to sonera.
10:22:55 <fizzie> Can't seem to find very specific information about the current situation here. From what I've seen, a single station can cover a reasonably large distance if you stick it on top of a fell, if you don't mind the fact that some of the valley-like parts get left out.
10:23:34 <oklodol> i do mind
10:24:21 -!- Slereah has joined.
10:24:23 <AnMaster> fizzie, I think large parts up there, are "nationalparker"
10:24:28 <AnMaster> not sure what that is in English
10:24:43 <oklodol> same
10:24:50 <fizzie> If you want to see Sonera's map, it's at http://mobileplaza.sonera.fi/matkapuhelin/kuuluvuus_kotimaassa.html -- it is clickable for closer look, and the colors for the detailed maps are: white = no-coverage, grey = needs-an-external-antenna, blue = normal GSM/GPRS, yellow = EDGE, green = 3G things.
10:25:29 <fizzie> Not too much white in there, but still some.
10:25:30 <AnMaster> protected areas of nature, you aren't allowed to do anything that will change it basically.
10:25:58 <AnMaster> fizzie, what is EDGE?
10:26:22 <fizzie> It's some sort of GPRS speed-upgrade, I think it was called "2.5G" somewhere too.
10:26:31 <fizzie> "Enhanced Data Rates for GSM Evolution".
10:26:46 <fizzie> http://en.wikipedia.org/wiki/EDGE
10:27:10 <AnMaster> mhm
10:27:18 <fizzie> "EDGE can carry data speeds up to 236.8 kbit/s --" though that's very much a theoretical thing only.
10:29:35 <fizzie> The 3G coverage is rather poor, though, pretty much cities only.
10:30:26 <fizzie> And special places; it seems they've stuck a single 3G base station at the Koli ski resort place, for example.
10:30:38 <AnMaster> heh
10:30:52 <AnMaster> is that some famous tourist attraction or such?
10:31:23 <fizzie> Well, maybe not "famous" but I guess it's reasonably popular.
10:31:51 <fizzie> Famous enough to have a Wikipedia article about it (well, not the ski resort, but the adjacent national park): http://en.wikipedia.org/wiki/Koli_National_Park
10:34:15 <AnMaster> ok
10:34:29 -!- ais523 has joined.
10:34:31 <AnMaster> oh so it is national park in English too?
10:34:33 <AnMaster> hi ais523
10:34:36 <ais523> hi
10:34:57 -!- GuestShadowSkunk has quit (Read error: 110 (Connection timed out)).
10:35:31 <ais523> and "national park" is certainly a concept in English
10:35:31 <ais523> although I can't tell if it's the same one you're referring to without more context
10:36:08 <fizzie> After the shutting-down of the ages-old 450 MHz analog NMT mobile phone network, they've reused that frequency range for the "@450" mobile broadband (1M/512k) which I guess has a better coverage than the 3G; 90 % of population, though that translates to some much lower percentage of geographical area.
10:37:06 <AnMaster> protected area of nature, you aren't allowed to do basically anything that can change it with some exceptions.
10:37:21 <fizzie> Maybe "ages-old" is a bit of an exaggeration: "The world's first NMT call was made in Tampere, Finland, in 1978.[1] The NMT network was opened in Sweden and Norway in 1981, and in Denmark and Finland in 1982."
10:38:12 <fizzie> If you want an "official definition" for a national park, I guess this listing suffices: http://en.wikipedia.org/wiki/World_Commission_on_Protected_Areas
10:38:20 <fizzie> Everyday use might differ, there you have to ask one of the natives.
10:45:01 -!- ais523 has quit (Read error: 104 (Connection reset by peer)).
10:46:18 -!- ais523 has joined.
10:49:17 -!- ais523 has quit (Read error: 104 (Connection reset by peer)).
10:49:29 -!- ais523 has joined.
10:50:19 <AnMaster> wb ais523
10:50:32 <ais523> thanks
10:50:39 <ais523> <IBM's lawyers> The term "concerning" shall mean relating to, referring to, reflecting, describing, evidencing, referencing, discussing or constituting.
10:52:25 <ais523> ooh, and Google are writing their own OS, using Linux's kernel (and presumably drivers, etc.) but none of the rest of the stuff that's normally used together with the kernel
10:53:07 <fizzie> If you logread, you'd know it was the original source for what led to the "national park" discussion.
10:53:11 <fizzie> But you don't. Woe is you!
10:53:26 <ais523> no, I don't normally have the focus to logread
10:53:40 <fizzie> If I were one of them IBM's lawyers, I'd add "concering" itself somewhere in the middle of that list.
10:53:46 <ais523> fizzie: haha
10:59:00 -!- ais523 has quit (Read error: 54 (Connection reset by peer)).
11:00:12 -!- ais523 has joined.
11:03:46 <AnMaster> wb ais523
11:05:17 <oklodol> AnMaster: oh so it is national park in English too? <<< what part of "same" did you not understand
11:05:31 <AnMaster> oklodol, where did you say it was the same?
11:05:44 <oklodol> 12:24… AnMaster: not sure what that is in English
11:05:45 <oklodol> 12:24… oklodol: same
11:06:26 <AnMaster> oklodol, ah, missed that
11:06:45 <fizzie> oklodol: You are so terse.
11:06:54 <fizzie> I didn't see the "same" either.
11:07:15 <fizzie> Why does that sound like an insult? "Ha ha, you're so terse!"
11:07:24 <oklodol> because terse also means stupid
11:07:38 <fizzie> Wordnet only lists: # S: (adj) crisp, curt, laconic, terse (brief and to the point; effectively cut short) "a crisp retort"; "a response so curt as to be almost rude"; "the laconic reply; `yes'"; "short and terse and easy to understand"
11:07:43 <fizzie> But I guess it could.
11:07:59 <oklodol> hmm.
11:08:15 <fizzie> "You're not one of those tersies, are you?"
11:08:19 <oklodol> at least something very close to it means stupid, or is used that way.
11:08:41 <oklodol> or your pun just deceived my brain
11:08:44 <oklodol> well
11:08:56 <oklodol> i guess it wouldn't have been a pun then
11:10:09 <oklodol> ah it's dense.
11:10:12 <oklodol> that took quite a while
11:10:37 <oklodol> possibly because i didn't think about it, but still
11:21:50 <fizzie> I had written that word already, then decided not to mention it for some reason.
11:22:07 <fizzie> Also I typoed "I had written that world already" first. Meaningful?
11:22:32 <oklodol> why would it be+
11:22:33 <oklodol> ?
11:38:10 -!- oerjan has joined.
11:38:52 <oklodol> hi oogie!
11:40:09 <oerjan> hi oclum
11:40:30 <oklodol> i'm feeling glemifandulous.
11:41:16 <oerjan> well that _does_ sort of sound like a problem with the brain, possibly the language center...
11:43:12 <oklodol> :DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD
11:43:21 <oklodol> superglio
11:44:27 <oerjan> http://en.wikipedia.org/wiki/Glial_cell
11:45:23 <oklodol> i'm not sure how that's helpful
11:45:43 <oklodol> i've told numerous times there's a well-known connection between all that is gli and all that is glue.
11:46:03 <ais523> okokokokokoko
11:46:04 <oerjan> yeah they sort of stick together
11:46:11 <oklodol> okokokokokokokokoko
11:46:14 <oklodol> oerjan: you're funny
11:46:18 <oklodol> okokokokokokokokokokokoko
11:46:21 <oklodol> okokokokokokokokoko
11:46:22 <oklodol> okokokokokokokoko
11:46:25 <oklodol> okokokokokokokokookokokokokoko
11:50:10 <oerjan> oklodol: that's easy when you have perfect straight men
11:50:28 <fizzie> O-cock-o.
11:50:36 <oerjan> rococo
11:51:54 <oklodol> i'm not sure i get it even with fizzie's explanaiton
11:51:56 <oklodol> *explanation
11:51:59 -!- BeholdMyGlory has joined.
11:52:16 <fizzie> What exploitation?
11:52:35 <oklodol> you should really buy those oculatory aids.
11:54:36 -!- dbc has quit (Client Quit).
11:57:34 -!- dbc has joined.
12:00:28 <oerjan> <fizzie> Also I typoed "I had written that world already" first. Meaningful?
12:00:33 <fizzie> Hah; since GPRS is called "2.5G" by some, and not everyone counts EDGE as "3G" since it's so sucky, some people apparently seriously refer to EDGE as a "2.75G" technology.
12:00:41 <oerjan> clearly you have subconscious delusions of grandeur
12:01:26 <fizzie> Where will it end? Soon there'll be a πG thing.
12:01:58 * oerjan assumes that character is a pi
12:02:35 <oerjan> for texnological reasons
12:02:38 <fizzie> Yes, that's what it should be.
12:03:14 <ais523> looks like a pi here
12:04:54 * oerjan is too lazy to go back to the logs to check right after he closed the page
12:05:50 <oerjan> and of course you all know already i'm too lazy to set up unicode correctly
12:16:51 -!- MigoMipo has quit ("QuitIRCServerException: MigoMipo disconnected from IRC Server").
12:38:01 -!- AnMaster has quit (Success).
12:38:27 -!- M0ny has quit.
12:43:25 -!- ais523 has quit (Remote closed the connection).
12:45:17 -!- oerjan has quit ("leaving").
12:45:19 -!- MizardX has quit ("What are you sinking about?").
13:06:21 -!- ais523 has joined.
13:14:01 -!- oklodol has changed nick to ogliopol.
13:33:15 <ogliopol> okay don't copy that was awesome :D
14:18:30 -!- upyr[emacs] has quit (Read error: 54 (Connection reset by peer)).
14:18:39 -!- upyr[emacs] has joined.
14:22:37 -!- Associ8or has quit ("#proglangdesign #ltu ##concurrency").
14:29:26 -!- upyr[emacs] has quit (Read error: 104 (Connection reset by peer)).
14:29:33 -!- upyr[emacs] has joined.
14:35:29 -!- MizardX has joined.
14:46:23 -!- MigoMipo has joined.
14:58:43 -!- MigoMipo has quit ("Page closed").
15:01:46 -!- Associat0r has joined.
15:25:27 -!- ais523 has quit (Read error: 104 (Connection reset by peer)).
15:37:25 -!- BeholdMyGlory has quit (Remote closed the connection).
15:51:17 <ogliopol> glia
15:51:41 <ogliopol> i should get lilja to start saying flia.
15:51:42 <ogliopol> *glia
15:53:28 -!- inurinternet has quit (Read error: 60 (Operation timed out)).
15:56:34 -!- ais523 has joined.
16:06:29 -!- darthnuri has quit (Connection timed out).
16:09:42 -!- inurinternet has joined.
16:09:45 -!- darthnuri has joined.
16:20:48 <ogliopol> o
16:20:48 <ogliopol> o
16:21:18 -!- inurinternet has quit (Client Quit).
16:24:26 <ais523> oko
16:24:45 <ogliopol> okokokokoko
16:24:58 <ogliopol> okokokokokokokokokokoko
16:25:10 <ogliopol> anyone speak italian here?
16:25:11 <ais523> hmm... I thought the next one would have 24 os
16:25:41 <ogliopol> 1, 1, 2, 6, 24?
16:25:46 <ogliopol> or just 2, 6, 24
16:33:53 -!- MizardX has quit (Read error: 54 (Connection reset by peer)).
16:34:00 -!- MizardX has joined.
17:12:09 -!- BeholdMyGlory has joined.
17:13:21 -!- MizardX- has joined.
17:13:26 -!- MizardX has quit (Connection reset by peer).
17:13:54 -!- MizardX- has changed nick to MizardX.
17:36:28 -!- ogliopol has quit (Read error: 110 (Connection timed out)).
17:47:00 -!- MizardX- has joined.
17:47:06 -!- MizardX has quit (Read error: 54 (Connection reset by peer)).
17:47:33 -!- MizardX- has changed nick to MizardX.
18:14:49 -!- coppro has joined.
18:21:09 -!- coppro has quit (Remote closed the connection).
18:21:58 -!- coppro has joined.
18:54:36 -!- darthnuri has quit (Read error: 60 (Operation timed out)).
18:59:13 -!- MizardX has quit (Connection timed out).
19:19:45 -!- olsner has joined.
19:25:16 <ehird> am i ehird or estoppel?
19:25:17 <ehird> ehird
19:25:21 <ais523> both
19:27:18 <ehird> 15:24:37 <pikhq> IE usage is now down to 56% overall.
19:27:28 <ehird> yeah suure
19:27:30 <ehird> what's the stats from
19:27:32 <ehird> firefox.com?
19:27:34 <ais523> ehird: it dropped 8% in a month, as a result nobody believes the statistics
19:27:36 <ehird> *where's
19:27:53 <ais523> and at least three of the major web stats people noticed it, but assumed it was something giving an incorrect reasing
19:27:54 <ais523> *reading
19:27:55 <ehird> ais523: the kind of websites to release such statistics tend to not be frequented by IE users
19:28:00 <ais523> and tbh, I think it's something like that too
19:28:01 <ehird> so I'd say it's bullshit
19:28:24 <ais523> a drop that sudden is unbelievable, something must have confused the browser detection
19:28:46 <ehird> ais523: google started running Chrome ads on TV in the usa...
19:28:50 <ehird> i just remembered
19:28:58 <ehird> sounds like we've found our culprit. maybe.
19:29:09 <ais523> yep, but firefox, chrome, and safari all stayed in proportion to each other
19:29:12 <ehird> i wondered why at the time, now it's obvious innit, chrome os
19:29:20 <ehird> ais523: hmm, true
19:29:47 <ais523> someone mentioned that that was about the release of IE8; IE8 taking 8% from IE7 is sort-of believable, and it could be that something's messed up in typical IE8 detection
19:30:28 <ehird> yay we have lambdabot now
19:30:34 <ehird> > 2+2
19:30:35 <lambdabot>   4
19:30:36 <thutubot>   4
19:30:36 <ehird> ^__^
19:30:46 * ehird wonders why thutubot is mirroring lambdabot
19:30:57 <ehird> 16:32:24 <ais523> AnMaster: someone's /msging thutubot
19:31:07 <ehird> ais523: thutubot is misinterpreting lambdabot's response as a message I think
19:31:28 <ehird> /msg doesn't even do anything
19:31:39 <ehird> 16:37:15 <ais523> it's meant to send the message to lambdabot, then relay the response
19:31:40 <ehird> 16:37:18 <ais523> it's broken, though
19:31:41 <ais523> ehird: it wasn't that, we found the correct answer eventually
19:31:41 <ehird> lol
19:31:50 <ehird> ais523: remove it then, so it doesn't make using lambdabot annoying?
19:31:58 <ehird> +haskell that is
19:32:02 <ehird> after all, it is just a cheap joke
19:32:04 <thutubot> but msging thutubot with instructions does reply to #esoteric
19:32:55 <ehird> 02:04:22 <AnMaster> So basically you can't use it where there is no internet connection I guess
19:32:55 <ehird> 02:07:15 <AnMaster> oklodol, yes right, but what if you are traveling and there happen to be no public wlan where you are or such.
19:33:05 <ehird> yeaaaaah in a few years everyone will have mobile broadband.
19:33:20 <ehird> designing something to account for people without the net is stupid, although I'm no fan of the cloud
19:34:09 <ais523> ehird: I don't even have a regular mobile phone
19:34:19 <ehird> ais523: mobile broadband has nothing to do with mobiles
19:34:24 <ehird> apart from usually being over the same network
19:34:31 <ais523> ehird: and both costing insanely large amounts of money
19:34:45 <ehird> ais523: you have a queer definition of insane
19:34:58 <ehird> it's some pounds a month
19:35:56 <ehird> 02:27:18 <fizzie> "EDGE can carry data speeds up to 236.8 kbit/s --" though that's very much a theoretical thing only.
19:36:04 <ehird> EDGE can carry 236kbits my ass.
19:36:07 <ehird> It's a bit better than dialup.
19:36:41 <ehird> 02:37:06 <AnMaster> protected area of nature, you aren't allowed to do basically anything that can change it with some exceptions.
19:36:43 <ehird> shit
19:36:46 <ehird> so nobody can walk in them?
19:36:58 <ehird> and it has to be stored at 0K
19:37:04 <ehird> damn you chaos theory! DAMN YOU ENTROPY!
19:37:32 <ehird> 02:53:26 <ais523> no, I don't normally have the focus to logread
19:37:36 <ehird> there was something about something
19:37:40 <ehird> that i directed at you if you logread
19:37:41 <ehird> lemme grep
19:37:50 <ehird> 11:50:31 <ehird> I wish ais523 logread.
19:37:50 <ehird> 11:50:36 <ehird> "One thing I found puzzling was that the Brits consistently apologized for and/or denigrated Birmingham."
19:37:52 <ehird> 11:50:41 <ehird> —Bruce Eckel, http://www.artima.com/weblogs/viewpost.jsp?thread=261930
19:38:48 -!- oerjan has joined.
19:38:54 <ehird> oklodol is so glio.
19:39:08 <oerjan> gliissimo
19:39:29 <ehird> 04:00:33 <fizzie> Hah; since GPRS is called "2.5G" by some, and not everyone counts EDGE as "3G" since it's so sucky, some people apparently seriously refer to EDGE as a "2.75G" technology.
19:39:35 <ehird> piG.
19:39:43 <ehird> It's marginally better than 3G, but costs pi more moneys a month.
19:39:51 <ehird> oh, fizzie said that
19:39:55 <ehird> damn logs corruptin'
19:40:21 <oerjan> better than exponential corruption
19:40:29 <ehird> HURF HURF DURF TURF
19:40:46 * oerjan does the heimlich maneuver on ehird 
19:40:55 <ehird> and you say you're not gay.
19:41:12 * oerjan wonders when he said that
19:41:18 <oerjan> or the opposite
19:41:41 <ehird> i should ask colin percival if i could get a bulk discount on tarsnap; i don't particularly feel like paying $696 a month if I feel like filling up my entire future 2.32TB harddrives with completely uncompressable, 100% random data
19:42:01 <ehird> PER BACKUP! PLUS BANDWIDTH!
19:42:04 <ehird> oh the huge manatee.
19:46:23 <oerjan> <ehird> designing something to account for people without the net is stupid, although I'm no fan of the cloud
19:46:39 <ehird> 19:46 oerjan: <ehird> designing something to account for people without the net is stupid, although I'm no fan of the cloud
19:46:40 <oerjan> you should account for temporary net loss, certainly
19:46:46 <ehird> of a few minutes.
19:47:07 <ehird> it's not unthinkable that in ten years, an area without internet would be considered dangerous
19:47:14 -!- AnMaster has joined.
19:48:18 <oerjan> few minutes?  when someone dug into a cable in this neighborhood it certainly took more than a few minutes to fix it...
19:48:53 <oerjan> (i assume they dug into a cable, someone was digging on the main road and internet disappeared, 2+2)
19:49:15 <ehird> we're more and more dependent on the interwebs. i truly think in a few years that long internet outages will become an unacceptable inconvenience and in a few more it'll be serious
19:49:44 <ehird> oerjan: pseudo-analogy: it takes more than a few minutes to fix cut power to a city, too.
19:49:57 <oerjan> sure
19:50:02 <ehird> does that mean we should design everything around not having power?
19:50:04 <ehird> i think not
19:50:09 <oerjan> hm good point
19:50:21 <pikhq> ehird: ... They're acceptable now?
19:50:39 <ehird> pikhq: heh :)
19:50:55 <ehird> pikhq: if all else fails, I'm sure we can send our packets up INTO SPACE
19:51:01 <ehird> HAVE YOU HEARD, PIKHQ HAS SATELLITE INTERNET! IN SPACE!
19:51:20 <oerjan> of course _big_ telecom disconnects get fixed fast.  my father who is a telecom engineer tells me the contracts include some hefty fines...
19:51:30 <ais523> ehird: netbooks are partly desired because of their long battery life
19:51:35 <ais523> which means, they are designed around not having power
19:51:41 <ehird> ais523: no they're not
19:51:45 <ehird> they're designed around having portable power
19:52:03 <ehird> guess what mobile internet-based devices are designed around?
19:52:05 <ehird> having portable internet.
19:52:26 -!- inurinternet has joined.
19:52:45 <ehird> portable inurinternet.
19:54:58 <oerjan> ais523: has anyone told graue about the undeletable spam?
19:55:16 <ehird> whoa
19:55:17 <ehird> clue me in
19:55:19 <ehird> what's the issue
19:55:28 <oerjan> (should that be undeletable or undeleteable?)
19:55:42 <oerjan> ehird: http://esoteric.voxelperfect.net/wiki/User:454_buy_viagra
19:55:45 <ehird> une deli table
19:55:47 <Deewiant> The former
19:55:50 <ehird> oerjan: how's it undeletable
19:56:02 <Deewiant> Or, if you prefer, indelible
19:56:05 <ais523> ehird: graue put up a spam filter against that sort of thing
19:56:12 <ehird> *facepalm*
19:56:13 <oerjan> any attempt to edit or delete it triggers the spam filter :D
19:56:17 <ais523> which means that nobody, not even admins, can do anything even referencing the page
19:56:21 <ais523> apart from look at it
19:56:24 <ehird> i might have some ideas
19:56:38 <ehird> ais523: mediawiki has an XML-RPC interface doesn't it?
19:56:46 <ais523> ehird: yes, but it isn't enabled on Esolang
19:56:48 <ais523> I thought of that too
19:56:50 <ehird> fuck!
19:57:13 <ehird> can i say, btw, that the spam filter is shit anyway?
19:57:16 <ehird> it never stops any spam.
19:57:39 <ehird> ais523: are any of the other APIs enabled?
19:57:58 <oerjan> how is the url escaping of characters again?
19:58:03 <ais523> not as far as I can tell; the one in core is disabled, and all the others are uninstalled extensions
19:58:03 <oerjan> %hex ?
19:58:03 <ehird> http://esoteric.voxelperfect.net/wiki/User:454_buy_viagra?action=edit&section=55
19:58:06 <ehird> ais523: you can edit it
19:58:07 <ais523> oerjan: %hex, 2 digits
19:58:07 <ehird> http://esoteric.voxelperfect.net/wiki/User:454_buy_viagra?action=edit&section=55
19:58:10 <ehird> just have to add a new section
19:58:13 <ais523> ehird: try previewing
19:58:16 <ehird> argh
19:58:22 <ais523> and saving comes to the same thing
19:58:35 <ais523> my guess is the filter blocks POST but not GET
19:58:39 <ais523> or maybe, a bit of both
19:58:47 <ais523> but it seems to block POST consistently, and GET only sometimes
19:58:48 <ehird> we should really move esolang off that crappy server with the crappy category policy. :P
19:58:55 <ehird> http://esoteric.voxelperfect.net/wiki/User:454_buy_viagra doesn't load any more
19:58:57 <ehird> whaddidido
19:59:08 <ehird> now it works but without c ss
19:59:09 <ehird> css
19:59:22 <oerjan> damn didn't help
19:59:43 <oerjan> the save removed the escape too
19:59:44 <ehird> hmm
19:59:49 <ehird> ais523: what actions can you do to pages again?
19:59:58 <oerjan> ehird: yeah i've seen that css thing
20:00:02 <ais523> off the top of my head, view raw edit delete protect
20:00:07 <ais523> oh, and history
20:00:08 <ehird> can you protect it
20:00:11 <ais523> probably a few more I've forgotten
20:00:19 <ais523> and I haven't tried protecting it, although that would be rather ridiculous
20:00:33 <ehird> just try it, hitting buttons won't do any harm at least
20:00:51 <oerjan> famous last words :D
20:01:05 <ais523> ehird: nope, same problem
20:01:12 <ais523> in fact, I can't follow any of the links from the protect page
20:01:14 <ehird> worst case, we lose a few copies of definitions, some literature on them, and a bunch of useless languages :D
20:01:18 <ais523> the referrer seems to be blocking things
20:01:22 <ehird> ais523: aha
20:01:24 <ehird> i was about to say that
20:01:29 <ehird> ais523: make a POST to delete it without any referer
20:01:34 <ehird> use a referer blocker or curl or something
20:01:46 <ais523> let's see if I can find a referer blocker Firefox extension
20:01:58 <Deewiant> ais523: refcontrol
20:02:04 <ehird> ais523: it'd be easier just to use curl imo
20:02:19 <ehird> ais523:
20:02:21 <ehird> ais523: http://cafe.elharo.com/privacy/privacy-tip-3-block-referer-headers-in-firefox/
20:02:28 <ehird> just set network.http.sendRefererHeader to 0
20:02:31 <ehird> assuming it's not out of date
20:02:48 <ais523> ehird: referers are helpful on certain other sites, though
20:02:53 <ehird> ais523: so turn it off afterwards
20:02:55 <ais523> a couple of my user scripts rely on them, for instance
20:02:56 <ehird> it is a one time thing...
20:03:12 <ais523> but a referer-blocker will be useful, esolang blocks all sorts of random referers for no good reason
20:03:20 <ehird> meh, ok
20:04:56 <ais523> "User:454 buy viagra" has been deleted. See deletion log for a record of recent deletions.
20:05:10 <ehird> \o/
20:05:15 <ehird> Precondition Failed
20:05:15 <ehird> The precondition on the request for the URL /wiki/Special:Recentchanges evaluated to false.
20:05:16 <ehird> lawl
20:05:25 <ais523> ok, now /that's/ bad
20:05:34 <ehird> that was just cuz i came from viagra
20:05:36 <ehird> errrrrrrrrrrrrrrrrrr
20:05:38 <ehird> reword that
20:05:49 <ais523> I don't have the perms to remove a deletion log entry
20:05:57 <ehird> no i mean
20:05:58 <ehird> referer
20:06:00 <ehird> was the problem
20:06:01 <ehird> for me
20:06:05 <ais523> yep, fair enough
20:06:06 <ehird> refreshing fixed i
20:06:06 <ehird> t
20:06:09 <ais523> good
20:06:19 <ehird> i think it's mod_security that's fucking up this stuff
20:06:24 <ehird> i remember on my old shared hosting
20:06:25 <ais523> what does mod_security do?
20:06:28 <ehird> you couldn't submit a form saying "Perl"
20:06:38 <ehird> because perl is sometimes used for programs which you could be attempting to inject, I think
20:06:47 <ehird> on the wiki I had I did P(bold, unbold)erl
20:06:48 <ehird> :D
20:06:51 <ehird> ais523: break things.
20:06:57 <ehird> in a feeble attempt at securing you.
20:07:03 <ais523> ehird: learn the <nowiki/> tag
20:07:12 <ehird> ais523: it wasn't mediawiki.
20:07:17 <ais523> oh, ok
20:07:26 <ais523> it's a neat no-op, it does the same thing as space in Perl or @ in Cyclexa
20:07:30 <ais523> *space in Perl6
20:07:46 <ehird> it's just /**/
20:07:49 <ais523> Cyclexa doesn't have an unspace to cancel it out, though, unless you're putting antitext in the original program
20:07:54 <pikhq> ehird: &#80;erl
20:08:00 <ais523> ehird: /**/ is equivalent to space by definition, at least in C
20:08:03 <ehird> pikhq: it filtered html out
20:08:12 <ehird> ais523: well, yes, but <nowiki/> is /**/
20:08:14 <ehird> for mediawiki
20:08:17 * oerjan bounces around in celebration
20:08:22 <pikhq> ehird: ... But not <b />?
20:08:36 <ehird> pikhq: (bold, unbold) is <b/>?
20:08:37 <ehird> news to me.
20:08:40 <ais523> pikhq: I'm guessing it was BBcode, which is not XML-like apart from having matching tags
20:08:45 <ehird> nah
20:08:45 <ais523> ehird: yes in XML, no in HTML
20:08:48 <ehird> it was like '''' or something
20:08:59 <ehird> lemme check
20:09:01 <oerjan> <ehird> reword that <-- no, let's not do that :D
20:09:03 <ehird> ais523: I meant, I never said <b/>
20:09:17 <ehird> ais523: it was bbcode, it turns out
20:09:21 <ehird> so P[b][/b]erl
20:09:35 <ehird> well, it links with [[foo]] for both external and internal, so not quite bbcode
20:09:43 <pikhq> That's pretty dumb.
20:10:27 <ehird> [[foo]] was just stolen from mediawiki because it's convenient, the rest was bbcode.
20:10:29 <ehird> i didn't make the software.
20:10:41 <ehird> ais523: why did wikipedia's nocover graphic change?
20:10:46 <ehird> licensing issues with a picture of a bloody cd case?
20:10:53 <ehird> please say no...
20:11:26 <ais523> ehird: I don't know the truth, although I'm guessing it's more likely an aesthetics edit war than licensing issues
20:11:40 <ais523> check the image history, it shouldn't be too hard to find
20:11:59 <ehird> ais523: it's a separate file; it'll be in the template
20:12:06 <ehird> ais523: but can you seriously argue http://en.wikipedia.org/wiki/File:Nocover.svg looks nicer than the older one?
20:12:23 <ehird> http://en.wikipedia.org/wiki/File:Nocover.png
20:12:24 <ehird> the old one
20:12:38 <ais523> ok, maybe it's an image format row
20:12:44 <ehird> haha
20:13:15 <ehird> wikipedia is so hopeless
20:13:22 <oerjan> Ei saa peittää
20:14:06 <ais523> I've seen what happened
20:14:17 <ehird> to what? wp?
20:14:19 <ais523> it is related to licensing issues, but not the ones you're thinking of
20:14:28 <ehird> wwwwwhaaaaaaaat
20:14:38 <ehird> Hopeless. Hopeless!
20:14:41 <ais523> basically, there were some rows that the "please upload a cover" image was turning up in cases where it was known that there were no free/fairable covers
20:14:53 <ehird> ARE YOU FUCKING SERIOUS
20:14:55 <ais523> so it was decided to leave the album cover section out altogether by default
20:14:57 <ehird> Jesus chris
20:14:57 <ehird> t
20:15:07 <ais523> now, it's been added back in on many pages
20:15:17 <ais523> but people are just specifying nocover.svg by hand, guessing the extension
20:15:24 <ais523> there are probably lots of others that use nocover.png
20:15:35 <ehird> wow, are you serious?
20:15:37 <ais523> it's a case of the template not specifying an image, so people have to guess
20:15:43 <ehird> what a load of bollocks
20:16:09 <ehird> i remember when you could use fair use images on the main page
20:16:12 <ehird> a good era that was
20:16:20 <ais523> ehird: http://en.wikipedia.org/w/index.php?title=Template:Infobox_Album&diff=196909019&oldid=192528717
20:16:40 <ehird> ais523: has it occurred to them to add a parameter |shownocover=no for those RARE cases?
20:16:46 <ehird> no. don't answer. obviously not.
20:16:59 <ehird> complete failure of rational thinking detected.
20:17:45 <ehird> ais523: ...answer. I'm curious.
20:18:39 <ais523> ehird: you'd get people shouting that it was nonstandardised if you did that
20:18:49 <ais523> (N.B. I am well aware of the contradictions involved here)
20:19:11 <ehird> do you think the court might make an exception for me if I give every one of these people a lobotomy?
20:19:15 <ehird> hypothetically, of course.
20:20:11 <pikhq> No jury would convict.
20:20:17 <ehird> hooray!
20:20:38 * ehird sees someone using #NUMBEROFEDITS-mod-N as a random number generator on wikipedia
20:20:47 <ehird> it's fast-paced enough that it could actually work...
20:21:41 <ehird> *NUMBEROFEDITS
20:21:57 <ehird> well
20:21:59 <ehird> it seems to be part of #expr
20:24:20 <ehird> ais523: "It is the very first time Microsofts monopoly over desktop computer OS market is seriously being challenged"
20:24:25 <ehird> [on chrome os]
20:24:28 <ehird> [talk:mainpage]
20:24:47 <ehird> pay no attention to the >10% market share in the corner.
20:31:00 <ehird> ais523: btw, the next firefox release (or whatever) will parse all html pages with the html 5 parser [you're highlighted because i last highlighted you and i remember you testing ubuntu releases]
20:31:06 <ehird> [tangential i know]
20:31:26 <ais523> ehird: it's backwards-compatible enough that it ought to work
20:31:36 <ais523> and I'll use Firefox 3.5 when it gets into ubuntu-proposed
20:31:42 <ehird> no, not firefox 3.5, I think
20:31:47 <ais523> because I'm not in the mood for fiddling with repos right now
20:31:47 <ehird> the wunafferthat
20:31:51 <ais523> ah, ok
20:31:54 <ehird> or maybe not, who knows?
20:32:00 <ehird> it's about:config in the nightlies, anyway
20:32:09 <ais523> I'll use that one when it gets to ubuntu-proposed too, if I'm still using that then
20:32:29 <ehird> http://ejohn.org/blog/html-5-parsing/
20:32:48 <ehird> ais523: nightly + set html5.enable to true, accurate as of yesterday
20:33:17 <ais523> ehird: I don't use nightlies of anything but TAEB and C-INTERCAL
20:33:37 <ais523> also, cfunge sometimes, but I hardly ever update that
20:33:38 <ehird> why not?
20:33:46 <ehird> don't say that, AnMaster will get angry
20:33:53 <ais523> and because I don't see the need, nightlies just change /too/ quickly for me
20:34:00 <ais523> and half the time I don't have an internet connection anyway
20:34:10 <AnMaster> ehird, why would that make me angry?
20:34:24 <ehird> AnMaster: you're always telling people to update cfunge or you'll go "la la la"
20:34:47 <ehird> ais523: oh, and a WTF from that article about the new parser: it's automatically translated from Java to C++
20:34:51 <AnMaster> ehird, about bugs yeah, it is a bit pointless to fix a bug that is already fixed in last version.
20:34:53 <ehird> because it comes from the validator.nu HTML5 parser
20:35:01 <ehird> they hand-wrote a Java→C++ translator just for it
20:35:08 <ehird> ais523: and the end result? 3% faster rendering. from translated java.
20:35:22 <ais523> 3% isn't a lot
20:35:37 <ais523> although is the Java → C++ translator still around/
20:35:40 <ais523> it could come in useful
20:35:43 <ehird> ais523: speed isn't the point; the point is that machine translating java to C++ somehow sped up hand-written C++ code
20:35:47 <ehird> and yes, but I think it's tailored just for the project
20:35:48 <AnMaster> ehird, every project tends to tell users to use last version when reporting a bug. Since often it is already fixed.
20:35:50 <ehird> they're going to keep it in-sync
20:36:11 <pikhq> Well, HTML5 parsing shouldn't break anything.
20:36:17 <ais523> AnMaster: I prefer people to report bugs in future versions before I've written them, to save me the testing trouble
20:36:21 <AnMaster> um. Why would auto-translated C++ be faster than hand written?
20:36:24 <ehird> ais523: it still has System.out.printnl and @Foo, i think they just make classes and #define stuff
20:36:28 <ehird> AnMaster: because the parser is different
20:36:37 <ehird> the point is that you'd think translating Java to C++ would result in non-optimal code
20:36:46 <pikhq> After all, most of the HTML5 parsing rules say exactly how you should parse malformed HTML.
20:36:55 <AnMaster> ehird, in theory you could hand write code equivalent of the converted one. Note "in theory".
20:37:06 <ehird> ... no shit
20:37:20 <AnMaster> (of course it might be practically infeasible, but since when have we cared about that in this channel)
20:37:30 * ehird dls firefox-3.6a1pre.en-US.mac.dmg
20:37:33 <ehird> AnMaster: the C++ is human-readable
20:37:43 <AnMaster> ehird, ok that is interesting
20:37:46 <ehird> i'm just saying that you'd expect the mismapping semantics to incur a performance penalty in general
20:37:51 <AnMaster> nothing like that generated by lex or yacc then
20:37:55 <ehird> so it's a surprise it ends up faster than the previous parser
20:38:00 <ehird> AnMaster: i doubt it works on all java code
20:38:08 <ehird> http://hg.mozilla.org/mozilla-central/rev/8b5103cb12a6
20:38:10 <ehird> here's the diff
20:38:11 <AnMaster> which is, (theoretically) human readable. But in practise not.
20:38:18 <ehird> (note: gigantic)
20:38:37 <ehird> it just looks like very boring, java-like C++ with some odd stuff that's #define'd away
20:38:52 <ehird> oh, and using the Java stdlib
20:38:54 <ehird> instead of the C++ one
20:38:59 <AnMaster> uhu
20:39:04 <pikhq> Which isn't hard to do with gcj.
20:39:11 <AnMaster> why not just compile the java to native code directly
20:39:14 <AnMaster> with gcj yeah
20:39:30 <ehird> AnMaster: because that (a) introduces a dependency on the system having a classpath (java stdlib implementation)
20:39:32 <pikhq> (gcj treats Java as a language with C++-like calling conventions, after all...)
20:39:35 <AnMaster> ah
20:39:38 <ehird> and (b) you can't call it easily from C++, prolly
20:39:43 <ehird> and (c) probably a bit slower, maybe.
20:40:00 <pikhq> ehird: Gcj-compiled Java is easy to call from C++.
20:40:03 <ais523> I like GCJ for the way it annoys Java purists
20:40:04 <ehird> OK, then
20:40:10 <ehird> but (a) is still the important thing
20:40:15 <AnMaster> but (d) it will probably crash less. Assuming gcj is managed and does bounds checks and so on.
20:40:23 <ehird> AnMaster: C++ can be managed.
20:40:38 <ehird> And since java's semantics don't map to C++, there'll be a library layer.
20:40:40 <ehird> Which is almost certainly managed.
20:40:45 <pikhq> GCJ-compiled Java can just about be called directly from C++.
20:40:48 <ehird> AnMaster: remember that any code that's good C is bad style in C++
20:40:57 <AnMaster> ehird, hah
20:41:01 <pikhq> (you have to include a GCJ library for the Java datatypes)
20:41:30 <ehird> pikhq: they're not going to introduce a compile-time dependency on gcj and they're not going to introduce a run-time dependency over gcj's runtime library
20:41:35 <AnMaster> can't you use gcj with something like -freestanding
20:41:37 <ehird> this is probably the sanest way to do it as far as mozille goes
20:41:39 <ehird> *mozilla
20:41:42 <ehird> AnMaster: the code uses java libs
20:41:47 <AnMaster> hm ok
20:41:47 <ehird> it is not mozilla-specific
20:41:57 <pikhq> ehird: I strongly suspect they already have a compile-time dependency on gcj.
20:42:08 <ehird> pikhq: o rly? the translator doesn't use gcj
20:42:09 <AnMaster> ehird, does this C++ code use a GC?
20:42:11 <ehird> and the pre-translated source is in-tree
20:42:18 <ehird> AnMaster: mozilla uses refcounting.
20:42:22 <ehird> XPCOM.
20:42:26 <ehird> (which is btw horrific)
20:42:46 <AnMaster> ehird, I thought mozilla was some mix of various in different parts?
20:42:48 <ehird> well, yes
20:42:49 <AnMaster> somewhat like memory management in GCC
20:42:51 <ehird> but it will be refcounted
20:43:07 * ehird replaces Minefield's unbelievably ugly icon with Firefox's
20:43:07 <AnMaster> sounds like a bad idea
20:43:11 <AnMaster> how to handle cycles?
20:43:21 <ehird> IT'S THE GLOBE WITH A CHEESY BOMB LOL
20:43:33 <ehird> AnMaster: i'm going to assume a fuckin' html 5 parser won't generate cyclic structures
20:43:42 <pikhq> ehird: It has an optional build-time dependency on a Java compiler.
20:44:06 <ehird> pikhq: keyword optional. but why?
20:44:14 <ehird> oh god, firefox's new logo is ugly
20:44:21 <ehird> the bits on the tail jus tlook all wrong
20:44:21 <pikhq> I haven't a clue.
20:44:25 <ehird> and the bit at the bottom looks like bad rescaling
20:45:06 <ehird> http://people.mozilla.com/~faaborg/files/shiretoko/firefoxIcon/firefox-64-noshadow.png
20:45:07 <ehird> just awful
20:46:25 <AnMaster> is refcounting ever a good idea for memory management? I guess there could be some specific cases...
20:47:37 <fizzie> A HTML parser can easily generate cyclic structures, if you want to have something like a DOM tree where you can go from an arbitrary node to its parent, and back.
20:48:19 <ehird> True.
20:48:24 <pikhq> AnMaster: If you don't have cyclic structures, it's not merely good, it's great.
20:50:15 <AnMaster> hm true
20:52:17 <fizzie> I'm not sure how incredibly great it is; there's a reasonable amount of work done for the reference-counting. Of course it's a tiny thing, but with some other GC styles you don't have to collect things very often either. But this is all just gut-feeling talk, I'm sure there's actual well-researched material written about this.
20:53:06 <ais523> <jeffb> (I note that the "process" of Slashdot incremental improvement has now reached a point where clicking anywhere in the text-entry box causes the box to LOSE focus. If you don't want us using Safari, there are more efficient ways to get us to move.)
20:53:14 <ais523> wow, Slashdot's UI gets worse by the day
20:53:18 <ehird> Doing something now is always worse than doing something later, performance-wise .
20:53:22 <ehird> s/ \././
20:53:37 <AnMaster> iirc python uses refcounting, that doesn't make sense though, unless it somehow can guarantee that the no cycles happen...
20:54:03 <ehird> AnMaster: it has a gc for cycles.
20:54:29 <fizzie> Perl uses refcounting too, and it just relies on the user to manually break cycles if they care enough. Though there's also a mark-and-sweep GC step done when the interpreter is being destroyed or something.
20:54:33 <AnMaster> ehird, um. how does that work??
20:54:35 <pikhq> The reference counting is just an optimisation to make the garbage collector's work easier.
20:54:52 <ehird> AnMaster: by confusing you.
20:55:17 <AnMaster> fizzie, interpreter being destroyed == /usr/bin/perl exiting? If so it seems a bit pointless.
20:55:27 <ehird> AnMaster: perl is more than perl(1).
20:55:32 <fizzie> AnMaster: It's for programs that embed a Perl interpreter for scripting purposes and such.
20:55:33 <ehird> it's a Perl interpreter library.
20:55:34 <ais523> AnMaster: thread ending also causes that
20:55:40 <ais523> in addition to what ehird's talking about
20:55:49 <ais523> and it matters due to DESTROY hooks even with one thread and no embedding
20:55:50 <ehird> oh jesus, firefox right-click menus still don't have curved edges like everything else because FIREFOX ON MAC DEVELOPERS HAVE NO BRAIN
20:56:07 <pikhq> ehird: ... Wow.
20:56:30 <pikhq> Why can't they just do what Qt does and use native widgets?
20:56:37 <ehird> pikhq: also the title/toolbar gradient still goes dark, light, dark
20:56:40 <ehird> also, qt fails on os x too
20:56:49 <ehird> there's something about the platform that causes everyone to lose their attention to detail
20:56:53 <ehird> i think it's the obnoxious users.
20:57:02 <ehird> pikhq: (as opposed to light, dark as normal os x)
20:57:08 <pikhq> That's because its OS X support is technically still a WIP.
20:57:15 <ehird> seriously, the time it took them to fake it with XUL? could have written a native cocoa interface.
20:57:43 <ehird> firefox's html5 parser seems to work great
20:57:56 <pikhq> Better than what they used to do; Qt used to just draw native-looking widgets.
20:58:02 <pikhq> Which is eeeew.
20:58:06 <ehird> pikhq: well that's the last time i used qt
20:58:08 <ehird> might be better now
20:58:16 <pikhq> Tried Qt 4?
20:58:27 <ehird> no, i'd like to
20:58:56 <pikhq> Qt 4 renders with native widgets whenever possible.
20:59:00 <ehird> good
20:59:22 <ehird> hey, now I'm using firefox I can... notice that its font-rendering still looks totally nonnative and ugly on a lot of stuff.
20:59:23 <ehird> sigh.
20:59:24 <ehird> but,
20:59:26 <ehird> mathml!
20:59:31 <ehird> html5 video!
20:59:36 <ehird> now i have to decide if it was worth that.
20:59:48 <ais523> ehird: but... Theora vs. h264
21:00:05 <ais523> Firefox on Mac OS X is in rather a weird place in that particular battle
21:00:14 <ehird> ais523: YOU SHOULD USE THEORA BECAUSE IF YOU USE THEORA THERE'S ABOUT A 30% CHANCE PEOPLE WON'T SUE YOU (DOWN FROM 0.01%)
21:00:19 <ehird> IF IT LOOKS BAD DRINK A FEW BEERS!
21:00:25 <ehird> Libre!
21:00:35 <ehird> *30% LESS CHANCE
21:00:52 <ehird> "You need the latest beta version of Firefox to view this demo!" —dailymotion.com's HTML 5 video test page
21:01:02 <ehird> i guess they removed it from the latest nightlies huh!
21:01:08 <ehird> oh, you mean they're sniffing versions?
21:01:09 <ais523> ehird: it's dailymotion is checking for a specific useragent string
21:01:09 <ehird> amazing!
21:01:12 <ehird> yes
21:01:14 <ehird> I was being sarcastic
21:01:17 <ais523> rather than using fallback like you're supposed to
21:01:36 * ehird tries a MathML test page; surely this will work?
21:01:40 <ais523> and acid3 tested fallback for <object> IIRC, so fallback for <video> may work as well
21:01:49 <ehird> wow the text rendering is so ugly.
21:02:26 <pikhq> Yes, fallback for <video> works.
21:02:49 <ehird> yay it works
21:02:52 <ehird>    
21:02:52 <ehird>     Bi
21:02:53 <ehird>     ⁡
21:02:55 <ehird>     (
21:02:57 <ehird>     x
21:02:59 <ehird>     )
21:03:01 <ehird>    
21:03:03 <ehird> ok, that's some seriously bad copy-paste behaviour
21:03:05 <ehird> for some notation that renderd as Bi(italics x)
21:03:07 <ais523> something went very wrong with that paste
21:03:17 <ehird> ais523: the source is:
21:03:19 <ehird> <math xmlns='http://www.w3.org/1998/Math/MathML'>
21:03:19 <ehird>    <mrow>
21:03:20 <ehird>     <mi>Bi</mi>
21:03:22 <ehird>     <mo>&#8289;</mo>
21:03:24 <ehird>     <mo>(</mo>
21:03:26 <ehird>     <mi>x</mi>
21:03:28 <ehird>     <mo>)</mo>
21:03:30 <ehird>    </mrow>
21:03:32 <ehird> </math>
21:03:34 <ehird> tl;dr browsers still don't get that you can't just convert shit to text to copy-paste it.
21:04:17 <ehird> wow, mathml doesn't have good spacing and text-sizing by default
21:04:24 <AnMaster> what browser is this?
21:04:30 <ehird> you'd think they'd not expect you to make a massive stylesheet just to get standard mathematical notation
21:04:33 <ehird> AnMaster: firefox nightly
21:04:36 <AnMaster> ah
21:04:37 <ehird> w/ html5 parser
21:04:49 -!- Gracenotes has quit (Connection reset by peer).
21:05:05 <ehird> data:text/xml;charset=utf-8,<math xmlns%3D'http%3A%2F%2Fwww.w3.org%2F1998%2FMath%2FMathML'>%0D%0A <mrow>%0D%0A  <munderover>%0D%0A   <mo>%26%238721%3B<%2Fmo>%0D%0A   <mrow>%0D%0A    <mi>m<%2Fmi>%0D%0A    <mo>%3D<%2Fmo>%0D%0A    <mn>1<%2Fmn>%0D%0A   <%2Fmrow>%0D%0A   <mn>3<%2Fmn>%0D%0A  <%2Fmunderover>%0D%0A  <mrow>%0D%0A   <munderover>%0D%0A    <mo>%26%238721%3B<%2Fmo>%0D%0A    <mrow>%0D%0A     <mi>n<%2Fmi>%0D%0A     <mo>%3D<%2Fmo>%0D%0A     <mn>1<%2Fmn>
21:05:07 <ehird> %0D%0A    <%2Fmrow>%0D%0A    <mi>m<%2Fmi>%0D%0A   <%2Fmunderover>%0D%0A   <mrow>%0D%0A    <mi>sin<%2Fmi>%0D%0A    <mo>%26%238289%3B<%2Fmo>%0D%0A    <mo>(<%2Fmo>%0D%0A    <mrow>%0D%0A     <msup>%0D%0A      <mi>x<%2Fmi>%0D%0A      <mi>m<%2Fmi>%0D%0A     <%2Fmsup>%0D%0A     <mo>%2B<%2Fmo>%0D%0A     <msup>%0D%0A      <mi>y<%2Fmi>%0D%0A      <mi>n<%2Fmi>%0D%0A     <%2Fmsup>%0D%0A    <%2Fmrow>%0D%0A    <mo>)<%2Fmo>%0D%0A   <%2Fmrow>%0D%0A  <%2Fmrow>%0D%0A <%2F
21:05:12 <ehird> mrow>%0D%0A<%2Fmath>
21:05:22 <ehird> ↑ the sum symbol should be bigger
21:05:23 <AnMaster> ehird, is it LR(1) to parse?
21:05:26 <ehird> and there should be spacing between the two sums
21:05:31 <ehird> AnMaster: nothing is.
21:05:39 <ehird> hmm depending on whitespace w/ irc sounds bad
21:05:41 <ehird> let's base64 it
21:05:42 <AnMaster> ehird, um... Python is even LL(1)
21:05:51 <ehird> AnMaster: python is adamantly simple to parse
21:05:55 <ehird> extensions are rejected because they break it
21:05:58 <ehird> data:text/xml;charset=utf-8;base64,PG1hdGggeG1sbnM9J2h0dHA6Ly93d3cudzMub3JnLzE5OTgvTWF0aC9NYXRoTUwnPg0KIDxtcm93Pg0KICA8bXVuZGVyb3Zlcj4NCiAgIDxtbz4mIzg3MjE7PC9tbz4NCiAgIDxtcm93Pg0KICAgIDxtaT5tPC9taT4NCiAgICA8bW8%2BPTwvbW8%2BDQogICAgPG1uPjE8L21uPg0KICAgPC9tcm93Pg0KICAgPG1uPjM8L21uPg0KICA8L211bmRlcm92ZXI%2BDQogIDxtcm93Pg0KICAgPG11bmRlcm92ZXI%2BDQogICAgPG1vPiYjODcyMTs8L21vPg0KICAgIDxtcm93Pg0KICAgICA8bWk%2BbjwvbWk%2BDQogICAgIDxtbz49PC9tbz4NCiAgICAgPG1uPjE8L21
21:06:00 <ehird> uPg0KICAgIDwvbXJvdz4NCiAgICA8bWk%2BbTwvbWk%2BDQogICA8L211bmRlcm92ZXI%2BDQogICA8bXJvdz4NCiAgICA8bWk%2Bc2luPC9taT4NCiAgICA8bW8%2BJiM4Mjg5OzwvbW8%2BDQogICAgPG1vPig8L21vPg0KICAgIDxtcm93Pg0KICAgICA8bXN1cD4NCiAgICAgIDxtaT54PC9taT4NCiAgICAgIDxtaT5tPC9taT4NCiAgICAgPC9tc3VwPg0KICAgICA8bW8%2BKzwvbW8%2BDQogICAgIDxtc3VwPg0KICAgICAgPG1pPnk8L21pPg0KICAgICAgPG1pPm48L21pPg0KICAgICA8L21zdXA%2BDQogICAgPC9tcm93Pg0KICAgIDxtbz4pPC9tbz4NCiAgIDwvbXJvdz4NCiAgPC9tcm93Pg0KIDwvbXJ
21:06:05 <ehird> vdz4NCjwvbWF0aD4%3D
21:06:13 <AnMaster> ehird, yes. Anyway saying nothing is LR(1) seems a bit wrong
21:06:22 <ehird> look up "hyperbole"
21:06:43 <AnMaster> ehird, is haskell LR(1)?
21:06:50 <ehird> I DON'T KNOW
21:06:54 <ehird> sheesh
21:06:57 <AnMaster> kay
21:07:08 <ais523> the whitespace would definitely need to be preprocessed first, in both Python and Haskell
21:07:20 <ehird> ais523: the lexer does that in python
21:07:26 <ehird> INDENT / DEDENT tokens
21:07:31 <ais523> makes sense
21:07:41 <ais523> so internally Python does have braces, it just doesn't let people use them
21:08:23 <ehird> bah, mathml is uselessly ugly in ff
21:08:34 <ehird> the sqrt sign can become disconnected from the top bar extending frmo it
21:08:37 <ehird> because it's just a border-top
21:08:43 <ehird> and pasting sucks
21:08:56 <ehird> guess I'll go back to my original plan: convert $whatever to TeX, render as png
21:09:06 <ehird> set alt to $whatever
21:09:43 <ehird> oh, and to boot, mathml has two formats
21:09:47 <ehird> semantics and presentation
21:09:52 <ehird> they're separate
21:09:58 <ehird> and if you just have semantics, firefox ignores all the tags
21:10:09 <ehird> so you need semantics+presentation or presentation+semantics
21:10:12 <ehird> only the latter works in ff
21:10:16 <ehird> the former behaves like just semantics
21:10:21 <ehird> and so everything is repeated twice!
21:16:09 <ehird> AnMaster: have you got the computer modern unicode fonts installed? http://canopus.iacp.dvo.ru/~panov/cm-unicode/
21:16:11 <ehird> i bet you have
21:16:42 <AnMaster> ehird, not sure. Since I have TexLive installed, I guess yes.
21:16:59 <ehird> AnMaster: texlive doesn't install system fonts i don't think
21:17:07 <AnMaster> ehird, ah you meant that way
21:17:08 <AnMaster> hm no idea
21:17:09 <AnMaster> maybe
21:17:13 <AnMaster> will look later, a bit busy
21:18:17 <oerjan> AnMaster: haskell's indentation sensitivity interacts with the parser in a more complicated way than python's, so you cannot even preprocess it.
21:18:58 <oerjan> so it's sort of like an LR(1) parser communicating both ways with indentation detection.
21:19:44 <oerjan> and that is in practice; the theoretical haskell definition has issues due to mixing operator precedence into this (that mixing will almost certainly be removed in the upcoming language revision)
21:20:14 <ehird> Latin Modern Roman, STIXNonUnicode, STIXSize1, STIXGeneral, Symbol, DejaVu Sans, Cambria Math
21:20:16 <ehird> it doesn't appear to work.
21:20:17 <ehird> oh well
21:20:34 <oerjan> (no(?) haskell compiler actually implements operator precedence as the standard says)
21:20:58 <oerjan> (wrt the insane interaction with indentation)
21:21:42 <ehird> it disappoints me that firefox still sucks on os x.
21:25:32 <oerjan> http://hackage.haskell.org/trac/haskell-prime/wiki/FixityResolution
21:26:56 <ehird> > \x -> x == x == True
21:26:57 <lambdabot>   Precedence parsing error
21:26:58 <lambdabot>      cannot mix `GHC.Classes.==' [infix 4] and `GH...
21:26:58 <thutubot>   Precedence parsing error
21:26:58 <thutubot>      cannot mix `GHC.Classes.==' [infix 4] and `GH...
21:27:01 <ehird> :-D
21:27:04 <ehird> ais523: can you shut thutubot up plz
21:27:06 <ehird> thx
21:27:17 <ais523> +quit
21:27:17 -!- thutubot has quit ("ThutuBot quitting").
21:27:20 <ais523> that'll do for now
21:28:33 <oerjan> that wouldn't type correctly even if parsed strictly according to the standard
21:28:37 <ehird> yes
21:28:41 <ehird> but it's still amusing
21:29:18 <ehird> "general.config.obscure_value;13" — about:config
21:29:27 <ehird> (I'm not joking)
21:29:42 <ehird> "The general.config.obscure_value preference specifies how the configuration file is obscured. Firefox expects that each byte in the file will be rotated by the specified value. The default value is 13. If this value is left unchanged, then the configuration file must be encoded as ROT13. Autoconfig will fail if the cfg file is not encoded as specified by this preference. A value of 0 indicates that the file is unencoded-- i.e. it is unobscured plain tex
21:29:44 <ehird> t. It is recommended that you set this value to 0. (This will allow you to skip the encoding step in part 3.)"
21:30:02 <oerjan> :D
21:30:19 <ehird> rot-13 is actually default
21:30:24 <ehird> now THERE'S a reason to use firefox!
21:30:27 <ais523> ok, the next question is: /why/ does Firefox have an option to rot13 the config file?
21:30:58 <ehird> ais523: to prevent grepping a large tree to find pseudo-sensitive data, presumably
21:31:02 <ehird> or for shits and giggles
21:31:12 <ehird> and to have a legit reason to include "general.config.obscure_value;13"
21:31:23 <ais523> ; or :
21:31:29 <ehird> ;
21:31:39 <ehird> if you copy an about:config entry it copies as tht
21:31:40 <ehird> that
21:31:43 <ehird> i dunno why
21:33:32 * ehird attempts to find a firefox extension that lets you view a page in the internet archive
21:34:04 <ehird> https://addons.mozilla.org/en-US/firefox/addon/2570 ← i'll have some of that please
21:34:24 <ehird> except it thinks it's for an older version
21:34:25 <ehird> grrrrrrrrrr
21:34:34 <ehird> mozilla.org, surely YOU don't force check user agent?
21:36:10 <ehird> yay, I can add it via another link
21:36:21 <ehird> Resurrect Pages 2.0.4 could not be installed because it is not compatible with Minefield 3.6a1pre.
21:36:22 <ehird> or not.
21:36:30 * ehird installs nightly tester tools
21:37:33 <ehird> ais523: apparently the firefox nightlies are built with gcc3……………………………
21:37:47 <pikhq> ehird: *Wut*?
21:37:51 <ehird> pikhq: yep
21:37:54 <ehird> nightly tester tools tells me so
21:38:00 <ehird> prolly the official builds are too
21:38:05 <ehird> mozilla architecture sucks so much
21:38:21 <pikhq> Odd; I build Firefox with GCC 4.
21:38:39 <pikhq> I also build it linking against XULrunner, which Mozilla doesn't support, so... Go Gentoo!
21:39:23 <ehird> Go please-will-gecko-die-and-give-way-to-webkit!
21:39:51 <pikhq> Or-at-least-a-complete-rewrite-of-Gecko!
21:39:56 <ehird>   Last Comment    Bug 262173 -  Firefox Icon Problem - new firefox icon appears to be giant red panda that is humping south america
21:39:56 <ehird> Summary: Firefox Icon Problem - new firefox icon appears to be giant red panda that is...
21:39:58 <ehird> https://bugzilla.mozilla.org/show_bug.cgi?id=262173
21:40:06 <SimonRC> lolol
21:40:11 <pikhq> Oh, wait. That would be written by Firefox devs.
21:40:12 <ehird> pikhq: webkit + html5 parser + more html5 element support = yum
21:40:16 <ehird> but, lol
21:40:22 <SimonRC> I wonder where the old mozilla ican was hiding?
21:40:31 <pikhq> Webkit FTW.
21:40:38 <pikhq> (or at least KHTML)
21:41:03 <ehird> pikhq: khtml? spoken like someone who's never used khtml for prolonged lengths of time!
21:41:12 <ehird> pikhq: btw you do know khtml is dead right?
21:41:29 <ehird> KDE is effectively switching/has switched to WebKit
21:41:32 <pikhq> Oh, KDE finally killed that and replaced it with QtWebkit?
21:41:37 <ehird> dunno if they have yet
21:41:39 <pikhq> Sweet.
21:41:41 <ehird> but they're certainly planning on it
21:41:48 <ehird> they might still call it KHTML but I doubt they'll change much
21:41:52 <pikhq> Makes sense; after all, WebKit is *part of Qt*.
21:42:09 <ehird> webkit also has a gtk backend, though
21:42:10 <ehird> and some others
21:42:12 <ehird> very nice
21:42:31 <ehird> konqueror doesn't really have any redeeming features imo
21:42:34 <ehird> as a web browser
21:42:44 <ehird> as a unified object-munger it's ok
21:42:58 <ehird> In November 2007, the project announced that it had accomplished support for HTML 5 media features, allowing for embedded video to be natively rendered and script-controlled in WebKit.[15]
21:43:02 <ehird> oh, so it has all that
21:43:05 <SimonRC> "unified object-munger"?
21:43:07 <ehird> guess it's just codec issues, not webkit
21:43:24 <ehird> SimonRC: with kparts and the uri stuff and shit, you can point konqueror at just about anything that can be described as a resource
21:43:29 <pikhq> Konqueror is strictly speaking not a web browser. It's just a shell for a bunch of KDE libraries.
21:43:42 <pikhq> It just so happens that you can use it as a web browser.
21:44:22 <SimonRC> ehird: so I can e.g. view a gif as a directory of pictures?
21:44:26 <SimonRC> *animated gif
21:44:32 <Deewiant> pikhq: It's like IE, then.
21:44:39 <SimonRC> or I can vaiew all the layers on a GIMP image?
21:44:44 <pikhq> Deewiant: Except with a significantly better design.
21:44:49 <ehird> SimonRC: no, it munges everything in the same way
21:44:54 <Deewiant> Probably, yes.
21:44:54 <ehird> it's a viewer, mainly
21:45:07 <SimonRC> but I can read a zip file with it?
21:45:17 <pikhq> Sure; that
21:45:21 <pikhq> 's the .zip KIO.
21:45:26 <ais523> even Firefox can read inside zipfiles
21:45:36 <ehird> konqueror's model isn't a good unified one
21:45:41 <pikhq> zip:///path/to/zip/
21:45:45 <ehird> you pretty much need the basic idea in an OS
21:45:56 <ehird> everything's an object, there aren't any applications, and you can munge and view it in various ways
21:45:59 <pikhq> The same URI works for all KDE apps.
21:46:04 <ehird> i know
21:46:12 <SimonRC> application stacks are thickening in two ways in these decades...
21:46:14 <pikhq> (unfortunately, KDE isn't OS-level, so that's not as nice as it could be)
21:46:22 <ehird> it's just too specific
21:46:31 <ehird> "you get to give us a URI and you get one way to view it"
21:46:41 <SimonRC> the web browser is turning into another layer, and virtualisation systems etc are turning into another full layer
21:46:46 <ehird> that's not "absolutely everything's an object and you can view and munge it and change this at runtime"
21:46:51 <ehird> SimonRC: yeah
21:47:01 <ehird> what i'm describing is basically smalltalk
21:47:09 <ehird> smalltalk was damn good shit
21:47:13 <SimonRC> but javascripty?
21:47:16 -!- KingOfKarlsruhe has joined.
21:47:19 <ehird> SimonRC: i'm talking about an os
21:47:23 <ehird> soooooooooooooooooo... no
21:47:52 <pikhq> Smalltalk: Objects done right.
21:48:15 <pikhq> (it is also the first time it was really done as a language feature, so it was first done right, and then other people fucked it up)
21:48:51 <ehird> given a large budget, lots and lots of of documentation material and a load of time and i'd be happy to code my ass off to make a next-generation smalltalk
21:48:55 <ehird> pikhq: smalltalk isn't a language
21:49:02 <ehird> the language is simplistic on purpose and entirely incidental
21:49:17 <ehird> smalltalk's an "environment"; which is what we call OSs without kernel code.
21:49:20 <ehird> (it was an OS originally.)
21:49:34 <SimonRC> ah, but will it run the thousands of essential legacy apps and games people want to use?
21:49:34 <pikhq> It's both a language and a runtime environment.
21:49:37 <ehird> the thing about a smalltalk os is that you have emacs basically built in
21:49:49 <ehird> SimonRC: fuck that shit! dual-boot windows like all of us gamers do.
21:49:54 <ehird> or use a vm.
21:50:00 <ehird> or just dammit i don't care if it's practical.
21:50:13 <ehird> pikhq: but the main thing is the OS. the language is so simple because it's not the important part.
21:50:28 <SimonRC> how about security?  HOw do you stop any object fucking with stuff it oughtn't?
21:50:47 <SimonRC> actually, that might not be too hard
21:50:47 <ehird> SimonRC: capability-based security, which comes out to ubiquitous, highly-fine-grained, total sandboxing
21:50:49 <SimonRC> hmm
21:50:57 <ehird> (and interestingly gives us an isomorphism for objects<->processes)
21:51:03 <pikhq> ehird: The language is still done right.
21:51:09 <ehird> pikhq: very true
21:51:16 <SimonRC> ehird: who has already done it?
21:51:31 <ehird> SimonRC: buncha research OSs so obscure I can't even name 'em
21:51:37 <ehird> it seems to Work.
21:51:39 <pikhq> SimonRC: ... This is all part of Smalltalk's design.
21:51:43 <ehird> pikhq: no
21:51:46 <ehird> capability security isn't
21:51:50 <pikhq> Oh.
21:51:54 <pikhq> Trivial to add on, though.
21:52:00 <ehird> SimonRC: http://www.eros-os.org/
21:52:01 <SimonRC> ehird: ask the VPRI for help ;-)
21:52:05 <ehird> http://www.capros.org/
21:52:07 <ehird> http://www.coyotos.org/
21:52:16 <SimonRC> they sound familiar
21:52:17 <ehird> http://en.wikipedia.org/wiki/E_programming_language
21:52:25 <ehird> SimonRC: coyotos is the reason for bitc
21:52:31 <ehird> or whatever that language is called
21:52:34 <ehird> c with scheme syntax, basically
21:52:53 <ehird> SimonRC: plash, that our EgoBot uses, is like primitive capabilities for linxu
21:52:54 <ehird> linux
21:53:14 <pikhq> Ah, bitc.
21:53:21 <pikhq> I was looking for that.
21:53:31 <ehird> SimonRC: apparently the famous L4 microkernel does some sort of capability stuff as seL4
21:53:33 <ehird> pikhq: it's dead
21:53:39 <ehird> pikhq: the main guy went to work for microsoft, on Singularity
21:53:45 <ehird> no wait
21:53:47 <ehird> pikhq: Midori
21:53:57 <ehird> pikhq: a managed-code operating system, iirc based on .NET (pikhq explodes into rage)
21:54:24 <ehird> Midori is the code name for a managed code operating system being developed by Microsoft Research. It has been reported[1][2]  to be a possible commercial implementation of the Singularity operating system, a research project started in 2003 to build a highly-dependable  operating system in which the kernel, device drivers, and applications are all written in managed code. It was designed for concurrency, and can run applications in multiple places.[3]
21:54:26 <ehird> It also features an entirely new security model that sandboxes applications for increased security
21:54:31 <ehird> <pikhq> BUT YOU CAN'T USE MANAGED CODE FOR AN OS! IT'S EEEEEEEVIL!
21:55:02 <pikhq> I am highly tempted to swear off C.
21:55:09 <ehird> Wow.
21:55:13 <ehird> How did _that_ happen?
21:55:33 <pikhq> I tried programming in it again.
21:55:39 <pikhq> Much anger.
21:55:42 <ehird> Ah. That usually does it.
21:56:33 <pikhq> Also discovering that well-written Haskell was plenty speedy.
21:56:54 <SimonRC> there is an all-haskell OS caled House
21:56:55 <ehird> (Caveat: For most things.)
21:56:59 <ehird> SimonRC: he knows I think
21:57:02 <ehird> there's also Iron or whatever
21:57:06 <pikhq> Well aware.
21:57:37 <SimonRC> putting apps in managed code is a nice idea
21:58:02 <ehird> my problem with abolishing C is that writing a JIT compiler purely in assembly for a high level language is fucking tedious
21:58:12 <SimonRC> once the theorem prover has shown it is safe, you don't need to worry about security so much, so you can get a speed boost
21:58:15 <SimonRC> hopefully
21:58:35 <SimonRC> ehird: do you mean JIT-to-C?
21:58:36 <pikhq> Just not from Microsoft. Please, not from MS.
21:58:47 <ehird> SimonRC: jit to machine code
21:58:57 <ehird> as in, since we can't use C, we need to implement our high-level language
21:59:05 <ehird> since we're doing an OS, it needs to be fast
21:59:12 <ehird> since we're doing runtime-modification, it needs to be a JIT or similar
21:59:18 <ehird> since we can't use C…
21:59:27 <ehird> we have to write our fast JIT for our high-level language in assembly
21:59:34 <ehird> at this point, I get bored and go do something on an existing OS.
21:59:41 * SimonRC is not sure about that at all
21:59:52 <ehird> SimonRC: what's your alternative?
22:00:11 <SimonRC> first, Singularity allows unsafe managed code in some bits of the kernel...
22:00:24 <ehird> SimonRC: that's not really related to what I'm saying
22:00:33 <SimonRC> secondly, you can use bootstrapping
22:00:42 <SimonRC> what are you saying then?
22:00:42 <ehird> SimonRC: i am wary of most bootstrapping.
22:00:50 <ehird> it scares me due to Reflections on Trusting Trust.
22:01:08 <ehird> i prefer the visible base case of another language as opposed to one sometime in the past when we made that binary
22:01:19 <ehird> SimonRC: i'm saying that, we've just booted up, initialized the gdt and shit
22:01:26 <ehird> we need to run our os code
22:01:29 <ehird> so we need our implementation
22:01:31 <ehird> and we can't use C
22:01:41 <ehird> so we gotta write our awesome, fast implementation of our nice language... in assembly.
22:01:45 <ehird> without even libc.
22:01:47 <ehird> or syscalls.
22:01:51 <ehird> just us, here. on the bare metal.
22:02:17 <SimonRC> um, why couldn't the assembly be assembled from a HLL?
22:02:27 <SimonRC> at the time that the boot image is created
22:03:01 <ehird> SimonRC: which? all ones other than the one we invented for the purpose suck, and do we really want to introduce a build dependency on a conventional OS?
22:03:08 <ehird> if we don't want to introduce that dependency, and allow building on our OS,
22:03:11 <ehird> we have to implement it TWICE
22:03:13 <ehird> and maintain it TWICE
22:03:16 <ehird> forever and ever
22:03:28 <SimonRC> why maintain it twice?
22:03:38 <ehird> SimonRC: so that people on non-this-OS systems can build it
22:03:58 <SimonRC> how does Linux get round this problem?
22:04:13 <ehird> SimonRC: by being a conventional os that works most the same as every other unix.
22:04:24 <ehird> allow me to consider that an unacceptable tradeoff...
22:04:48 <pikhq> Why can't the assembly be created from a HLL that's implemented in your nice language?
22:04:52 <pikhq> :P
22:05:02 <ehird> because 22:01 ehird: i prefer the visible base case of another language as opposed to one sometime in the past when we made that binary
22:05:16 <ehird> the OS should be recreatable from the source and a conventional OS
22:05:53 <SimonRC> but once MS releases 1000000000 copies, it is a conventional OS
22:06:11 <SimonRC> plus, being writen for managed code will make the compiler quite portable
22:06:11 <pikhq> Say, use Haskell. Write a HLL in Haskell that targets asm, and write a runtime environment in that.
22:06:18 <ehird> SimonRC: i know, the popularity argument
22:06:22 <ehird> it's a way of getting away with it
22:06:23 <ehird> not a justification
22:06:40 <ehird> pikhq: my haskell os involved carefully writing non-consing haskell code to build up a memory allocation system
22:06:45 <ehird> (it didn't work; GHC always conses.)
22:06:52 <SimonRC> you like C because it has 1000 implementations then?
22:06:56 <pikhq> ehird: You actually tried writing one?
22:06:57 <ehird> SimonRC: who said I like C?
22:07:04 <ehird> pikhq: yeah... for about 50 lines.
22:07:08 <ehird> ;-)
22:07:12 <pikhq> Ah.
22:07:19 <ehird> pikhq: i was obeying the "no C, no POSIX" rule
22:07:29 <ehird> so I couldn't use the extract-shit-from-your-system-to-make-ghc-binaries-work thing
22:07:32 <pikhq> Clearly, the runtime environment should be Brainfuck.
22:07:37 <ehird> well ok just enough POSIX to make ghc work
22:07:40 <ehird> and I didn't want to patch ghc like house
22:07:41 <pikhq> That's easy to implement in asm.
22:07:52 <ehird> so I had to write all the stuff ghc uses in asm and non-consing haskell
22:08:04 <SimonRC> "non-consing Haskell"?
22:08:09 <ehird> SimonRC: cons = allocate
22:08:14 <SimonRC> I know
22:08:18 <SimonRC> sounds spectacular
22:08:26 <SimonRC> sounds bloody impossiblwe
22:08:43 <ehird> SimonRC: (# it#, unsafeDoesntWork# (evenIfYou# +# addHashesAndUnsafeNessBangPatternsToEverything#) #)
22:08:45 <pikhq> ehird: Maybe use a different systems programming language?
22:08:51 <ehird> but you can get quite close
22:09:04 <ehird> pikhq: like what? i'd only use c due to ubiquity, the other systems programming languages suck because they're not high-level :P
22:09:20 <ehird> anyway, if I was gonna write a new-smalltalk-OS, I'd just use C
22:09:24 <ehird> for the base of the kernel
22:09:27 <ehird> and the JIT
22:09:38 <ehird> get the gdts up and stuff, jump into the jit
22:09:40 <SimonRC> like squeak does?
22:09:50 <ehird> SimonRC: sorta
22:10:01 <ehird> the problem is that I don't want to depend on C code at runtime, as such
22:10:10 <ehird> with a C JIT, you're always using C code while using the OS
22:10:11 <ehird> always
22:10:16 <SimonRC> aren't JITs quite complicated programs?  I expect MM will be a bugger
22:10:28 <ehird> SimonRC: they're not that complicated... but good ones are.
22:10:35 <ehird> memory management, welp, reserve a pool
22:10:43 <pikhq> ehird: C or write one yourself.
22:10:56 <ehird> bah
22:11:02 <ehird> the clear long-term solution is an asm jit
22:11:07 <ehird> but i'm just not masochistic enough :)
22:11:09 <SimonRC> why?
22:11:29 <ehird> SimonRC: it's bootstrapless and dependenciless
22:11:40 <SimonRC> ASM requires computer work to get executable code just like a HLL?
22:11:44 <SimonRC> where is the advantage?
22:11:56 <ehird> eh
22:12:58 <SimonRC> there isn't much quantitative difference between asm and hll is there?
22:13:10 <SimonRC> except for eas of use and resource reqs
22:13:16 <SimonRC> *qualitative
22:14:54 <pikhq> SimonRC: And that an assembler is trivial to implement.
22:15:02 <SimonRC> hm, true
22:15:19 <SimonRC> in that case it might be worth going for a Forth though
22:15:28 <pikhq> Oh, likely.
22:15:37 <SimonRC> which will make the assembler easier to implement too
22:15:38 <pikhq> Much nicer to write for.
22:20:14 * SimonRC goes
22:21:39 * SimonRC goes
22:25:32 <ehird> back
22:25:39 <ehird> eh
22:25:46 <ehird> forths, meh, not close enough to the metal
22:26:38 <pikhq> You can peek and poke. how much closer do you want? :P
22:27:33 <pikhq> (yeah, yeah, that's about as silly as doing systems programming in BASIC)
22:28:05 <ehird> you can peek and poke via an interpreter
22:29:52 -!- ais523 has quit (Remote closed the connection).
22:36:30 -!- Pthing has quit (Remote closed the connection).
23:25:53 <lifthrasiir> anyone interested in this: http://cosmic.mearie.org/2009/07/tmax-window/ ?
23:26:47 <ehird> it looks like a very obscure company's rubbish nonsense vaporware, but I like reading in-depth about things that don't matter, so :D
23:27:32 <ehird> they like microkernels. why am i surprised?
23:28:06 <oerjan> mu.
23:28:19 <lifthrasiir> they also made own office suites and browser. duh.
23:28:58 <ehird> lifthrasiir: what's that in reply to?
23:29:19 <lifthrasiir> what do you mean?
23:29:51 <ehird> 23:28 lifthrasiir: they also made own office suites and browser. duh.
23:29:53 <ehird> what's the "duh" meaning?
23:29:56 <ehird> i don't see the contet
23:29:57 <ehird> context
23:30:14 <lifthrasiir> ah... it'd be "huh". hmm.
23:30:55 <ehird> ah
23:30:59 <lifthrasiir> as I wrote in the article (yes i wrote it) the *own* office suites and browser are based on oo.o and webkit respectively. :S
23:31:03 <ehird> "duh." as a statement normally means "obviously"
23:31:10 <ehird> lifthrasiir: i figured it was you from the domain
23:31:26 <lifthrasiir> as you saw domains like hg.mearie.org, reasonable.
23:31:42 <ehird> 23:30 Twey: Mr. Wing, sir, if you are listening: you have the I.Q. of a semolina pudding
23:31:42 <ehird> 23:30 Twey: And your homepage is full of <font> tags
23:31:44 <ehird> 23:31 Twey: I'm not sure which is the greater insult, but at least one is objectively true
23:31:45 <ehird> —#haskell
23:33:14 <lifthrasiir> someone said "Office and Scoutor is good because OO.o and Webkit is good. and since Window is not good it cannot be a fake." lol.
23:34:35 <ehird> lol
23:34:49 <ehird> "# If TmaxSoft used Wine or ReactOS to meet the deadline, they should release its source code in same license. But there was an instance that TmaxSoft didn’t respect other’s copyright.6 Of course this would not be a problem if they wrote the code from scratch."
23:34:59 <ehird> isn't the korean software industry kind of full of knockoff copyright infringements?
23:35:00 <ehird> no offense
23:35:06 <ehird> it just seems that way a lot
23:35:11 <ehird> ow, their UI is ugly
23:35:18 <ehird> windows 95 with an ugly paintjob on the borders
23:35:30 <lifthrasiir> ehird: for GPL violation, yes.
23:44:34 <ehird> [[    We are now building the perfect OS with UIs of Mac OS X, portability of Linux and compatibility with Windows.]]
23:44:42 <ehird> looking at their screenshot, somehow I doubt that on the UI front :)
23:44:57 <ehird> The official announcement was on July 7, 2009, in Grand Intercontinental Seoul. This was also the TmaxDay 2009, and entirely dedicated to Tmax Window, Office and Scoutor. The event took place in Grand Ballroom capable for 1,800 persons. But there were at least estimated 2,500–3,000 persons came to the event, so remaining people was watching big screen in the front of ballroom. Separate outdoor event was also scheduled but canceled due to rain.
23:45:02 <ehird> ok, talk about extravagant
23:45:51 * oerjan suggests the name HubrOS
23:46:11 <oerjan> or perhaps Hubrix
23:46:14 <ehird> it can go with the KDE-running-on-JSMIPS project, MolassOS
23:46:43 <ehird> [[Word 2003 ran flawlessly but its splash screen was broken (failing to render copyright texts).]]
23:46:47 <ehird> lifthrasiir: sure that wasn't intentional? :D
23:47:14 <lifthrasiir> ehird: i don't know of course. :p
23:48:08 <lifthrasiir> "Its user experience is modeled after Microsoft Windows due to its user friendliness, and microkernel is chosen for its high stability. (The presenter however also said that they are going to make another distinct user experience by the end of this year.)"
23:48:12 <lifthrasiir> (about UIs)
23:48:25 <ehird> windows is sooo user friendly
23:48:29 <ehird> i wonder where that myth came from.
23:48:55 <ehird> every tech literate and illiterate i've converted to ubuntu found it a breeze while windows confused or annoyed them
23:49:48 <lifthrasiir> ehird: if consider about Korean environment, Linux is not yet user friendly as there are full of English :p
23:49:51 <lifthrasiir> considering*
23:49:55 <ehird> that's true
23:50:11 <lifthrasiir> still i cannot agree on window's "user friendliness".
23:50:23 <ehird> *windows' btw :p
23:50:30 <lifthrasiir> ooh :p
23:52:21 <ehird>     First let’s view [Acid3 test page] with Internet Explorer. (Launches Internet Explorer 8) Okay, it failed with 12 [out of 100] points and… Shall we wait more? …Yes it scored 20. Then let’s view same page [with Scoutor]. (Launches Tmax Scoutor, and scores 98 out of 100) Oh… Let’s try once more. (Screen reloaded, in this time scores 99 out of 100) Ha, we made it to 99 points. (Applause)
23:52:37 <ehird> lifthrasiir: how the hell did they make the browser's rendering quality nondeterministic?!
23:52:48 <ehird> smoothness results aren't part of the score iirc
23:54:46 <ehird> "You see Flash-based video is played with Adobe Flash plugin."
23:54:51 <ehird> as opposed to playing them with unicorns
23:55:23 <ehird> "I remind that the female [after the computer]"
23:55:36 <ehird> lifthrasiir: does that sound as objectifying as "the female" does in english? :-D
23:58:09 <ehird> "Eh, the balance is only 100 won, it was originally much but it seems all drawn." ← 7 cents, this company is rich :D
23:58:59 <ehird> lifthrasiir: are living costs lower in kr?
23:59:11 <ehird> i mean i can't imagine anyone saying "we only have 7 cents here"
23:59:14 <ehird> vs "oh, it's empty"