01:07
<GPHemsley>
hsivonen: Hey, your validator sploded on me
01:08
<GPHemsley>
"Internal Error: Oops. That was not supposed to happen. A bug manifested itself in the application internals. Unable to continue. Sorry. The admin was notified."
01:34
<MikeSmith>
well, that's weird
01:34
<MikeSmith>
the "html5" twitter account quit working around end of September
01:34
<Hixie>
wohay, someone finally noticed!
01:34
<MikeSmith>
because I do not have it set up to do Oauth
01:35
<Hixie>
you are the first person to notice
01:35
<MikeSmith>
well, I noticed
01:35
<Hixie>
well, second
01:35
<Hixie>
i noticed when it broke
01:35
<Hixie>
(it=the whatwg twitter stuff)
01:35
<Hixie>
and didn't do anything about it because nobody complained
01:35
<MikeSmith>
Hixie: yeah I mean the "html5" one -- I guess the whatwg one is the same thing
01:35
<Hixie>
yup
01:36
<Hixie>
i took it as a sign that twitter was pointless
01:36
<MikeSmith>
I think I actually just copied your script when I set up the "html5" one
01:36
<MikeSmith>
heh
01:36
<MikeSmith>
amen
01:36
<MikeSmith>
for this kind of stuff at least
01:36
<Hixie>
all i've ever seen on twitter is people bitching
01:37
<MikeSmith>
twitter is now my means for exploring languages I don't speak
01:37
<MikeSmith>
it's useful for that at least
01:37
<MikeSmith>
and I've switched to only tweeting in languages that I don't actually speak nor understand well yet
01:37
<Hixie>
heh
01:38
Philip`
encountered http://twitlonger.com/ today, which seems quite absurd
01:38
<MikeSmith>
I figure, if I'm going to tweet, I might as well learn a little about it each tiem
01:38
<MikeSmith>
Hixie: anyway, what I was saying was weird was that an "html5" tweet did just get posted
01:39
<MikeSmith>
in spite of me doing nothing to "fix" the no-oauth "problem"
01:39
<MikeSmith>
http://twitter.com/#!/html5/status/28355466520
01:39
<Hixie>
um
01:39
<Hixie>
odd
01:39
<MikeSmith>
so, I don't know why that one went through
01:39
<MikeSmith>
not that I care so much
01:39
<MikeSmith>
just… odd
01:39
<Hixie>
o_O
01:39
<Hixie>
that _is_ odd
01:40
<MikeSmith>
Philip`: the better way to twit longer is to use Chinese
01:41
<MikeSmith>
maybe you heard me say this already, but if you tweet in English, it takes 2.5 to 3 times as much characters as when you tweet in Chinese
01:41
<Hixie>
the better way to tweet longer is to use buzz
01:42
<MikeSmith>
what's buzz? :)
01:43
<MikeSmith>
Philip`: what does http://www.twitlonger.com/ actually do?
01:43
<MikeSmith>
posts multiple tweets in serial?
01:44
<Philip`>
It posts the first n characters of your message and then a link to a page that has the rest of it
01:46
<MikeSmith>
ah
01:47
<MikeSmith>
well, big whoop
01:47
<MikeSmith>
site should then be called "howtotwitlongerforthetotallyclueless"
01:47
<MikeSmith>
or something
01:47
<MikeSmith>
but I guess that was your point
01:48
<MikeSmith>
absure
01:48
<MikeSmith>
absurd
01:48
<MikeSmith>
dammit
01:48
Philip`
always bears in mind the words of our now Prime Minister, "The trouble with Twitter, the instantness of it - too many twits might make a twat."
01:48
<MikeSmith>
I almost always type "absure" instead of "absurd"
01:48
<MikeSmith>
heh
01:48
<MikeSmith>
did dude really say that?!
01:48
<MikeSmith>
that is gold
01:49
<Philip`>
http://news.bbc.co.uk/1/hi/8176194.stm
01:51
<MikeSmith>
absurd is "hurt" in Welsh
01:52
<MikeSmith>
järjetön
01:52
<MikeSmith>
in Finnish
01:52
<MikeSmith>
which doesn't quite meet my "easier to type" requirement
01:53
<MikeSmith>
there's no really good recognizable cognate for "absurd"
01:53
<MikeSmith>
except for "absurd" that is i
01:54
<MikeSmith>
most European languages just seem to have some variation of "absurd" or "absurd" itself
01:54
<MikeSmith>
I guess because it came from Latin
01:55
<MikeSmith>
RIDICULARIUS
01:55
<MikeSmith>
now, that's a great word
01:55
<MikeSmith>
way better than just plan "ridiculous"
02:00
<MikeSmith>
"surreal" is "swrrealaidd" in Welsh
02:00
<MikeSmith>
even _looks_ surreal
02:00
<MikeSmith>
Welsh is a trip
02:03
<MikeSmith>
Google Translate gives "epätodellinen" for Finnish translation of "surreal"
02:03
<MikeSmith>
which actually seems to be more like "unreal"
02:03
<MikeSmith>
http://en.wiktionary.org/wiki/ep%C3%A4todellinen
02:03
<MikeSmith>
epä- +‎ todellinen
02:04
<MikeSmith>
epä- = "un-"
02:07
MikeSmith
wonders if gsnedders speaks Scottish Gaelic at all
02:08
<MikeSmith>
Wiktionary is a great source for examining translations for single words
02:08
<MikeSmith>
e.g., http://en.wiktionary.org/wiki/surreal
02:09
<MikeSmith>
but you need to set the Visibility > Show translations option
02:10
<MikeSmith>
http://en.wiktionary.org/wiki/os-fh%C3%ACreach#Scottish_Gaelic
02:10
<MikeSmith>
os-fhìreach = surreal
02:11
MikeSmith
shuts up because he realizes that the odds of anybody else here caring to hear about this here are close to nil
03:07
<robman>
afternoon all
03:09
<MikeSmith>
hey robman
03:09
<MikeSmith>
how's the weather in Australia?
03:09
MikeSmith
misses it already
03:09
<robman>
sunny and hot
03:09
<MikeSmith>
nice
03:09
<robman>
you back in japan?
03:09
<MikeSmith>
yup
03:09
<robman>
noice
03:09
<MikeSmith>
at home but heading out for lunch soon
03:10
<MikeSmith>
with a friend here who works for Mozilla
03:10
<robman>
yum...have fun
03:10
<MikeSmith>
so I got to schlep off to the other side of town
03:10
<MikeSmith>
I wish they'd move their office somewhere more convenient for me :)
03:11
<robman>
schlep is a word that's appearing in my life more and more nowadays
03:11
<MikeSmith>
heh
03:11
<MikeSmith>
well, that sounds bad
03:12
<robman>
i was lugging way too many laptops and a tripod all over Seoul - I feel shorter now
03:12
<MikeSmith>
ah yeah
03:12
<MikeSmith>
how did that meeting go, btw?
03:12
<MikeSmith>
the AR thing, right?
03:12
<robman>
really productive
03:12
<robman>
yah
03:12
<robman>
very intereting bunch of people
03:12
<MikeSmith>
ETRI people were there?
03:12
<robman>
yos
03:12
<robman>
and ogc
03:13
<MikeSmith>
great
03:13
<robman>
and khronos
03:13
<robman>
etc
03:13
<MikeSmith>
oh
03:13
<MikeSmith>
Khronos is a big plus for sure
03:13
<robman>
Neil was awesome...he really listened and engaged
03:13
<MikeSmith>
great
03:13
<robman>
wants to evangelise a common time system across all sensor related things to the khronos members
03:13
<robman>
would be great for sensor fusion/etc
03:14
<MikeSmith>
interesting
03:14
<MikeSmith>
did WebGL come up at all?
03:14
<robman>
aboslutely
03:14
<MikeSmith>
neat
03:14
<MikeSmith>
minutes available anywhere yet?
03:14
<robman>
and he's really keen to work with the w3c folks on improving the composition model
03:14
<robman>
yeah...holon
03:15
<MikeSmith>
what would "improving the composition model" mean in practice, I wonder?
03:15
<robman>
general notes here http://www.perey.com/ARStandardsMeetingOutputs.html and his notes here http://www.perey.com/ARStandards/Khronos_Standards_for_Mobile_AR.pdf
03:15
MikeSmith
takes a gander
03:17
<robman>
MikeSmith when you get a chance please checkout this too http://www.slideshare.net/robman/patterns-of-interest-iswar-seoul-2010
03:17
<MikeSmith>
will do
03:17
<robman>
i'd especially like to discuss slides 18-20 with you 8)
03:17
<MikeSmith>
hai
03:18
<MikeSmith>
man
03:18
<MikeSmith>
this whole presentation looks pretty interesting actually
03:18
<robman>
aw shux
03:19
<MikeSmith>
I have personally been away from this stuff for a while
03:19
<MikeSmith>
professionally I mean
03:19
<MikeSmith>
I don't remember if I told you that before W3C I did work mostly for the mobile market
03:19
<MikeSmith>
product dev
03:19
<MikeSmith>
backend stuff
03:19
<MikeSmith>
and client stuff
03:19
<MikeSmith>
also
03:19
<robman>
cool
03:20
<robman>
i gotta run to a meeting...have a great lunch and lets chat later
03:20
<robman>
boi
03:20
<MikeSmith>
in your slides, what are "triggers"?
03:21
<MikeSmith>
eh?
03:21
<MikeSmith>
I don't know "boi" either
03:21
<MikeSmith>
acronym?
03:22
<robman>
sorry...that was bye
03:22
<robman>
8)
03:22
<robman>
ill bbak later
03:23
<MikeSmith>
k
03:23
<MikeSmith>
I gotta go soon too
03:24
<MikeSmith>
but wanted to say you might want to consider coming to the Internet of Things conference in Tokyo at the end of November
03:24
<MikeSmith>
I helped out with review of some papers for that
03:25
karlcow
wishes he could go to this conf, but can't go out of Canada before Jan
03:25
<MikeSmith>
too bad for you karlcow
03:25
<MikeSmith>
I hope you learned your lesson
03:25
<karlcow>
living in Japan too long? :p
03:26
<MikeSmith>
I meant how you tried to smuggle your crack pipe through customs
03:27
<karlcow>
hehe. himitsu
04:47
<wirepair>
i am amazed at how many parsers insecurely handle external xml entities
04:47
wirepair
just found *another* bug in an RSS reader...
08:46
<zcorpan>
Hixie: maybe ie's way of dealing with <meta http-equiv=x-ua-compatible> is screwing you over in your navigation tests?
09:05
<Hixie>
zcorpan: how so?
09:06
<Hixie>
i mean, i don't have it set
09:07
<Hixie>
and i don't see why what i'm testing would be relevant to that
09:07
<Hixie>
surely it'll be the same in all the modes...
09:09
<zcorpan>
Hixie: i don't know how it works, but maybe ie is first parsing looking for the meta, and then reparsing if it didn't find one
09:10
<Hixie>
i doubt it, that'd be a perf nightmare
09:10
<Hixie>
also the timings don't fit that hypothesis (some of the tests have timestamps)
09:11
<zcorpan>
ok
09:12
<zcorpan>
(iirc ie reparses when turning on designMode, which is also perf nightmare but doesn't happen as often)
09:26
<zcorpan>
https://developer.mozilla.org/en/JavaScript_typed_arrays - does gecko support binary messages in websockets?
09:29
<annevk>
I don't think so
09:36
<zcorpan>
http://yfrog.com/f/2fejgrj/
09:37
<annevk>
old news
09:37
<annevk>
but still funny
09:47
<annevk>
jgraham, btw, I meant "WHATTF" last night
09:47
<annevk>
not sure about action man and bureaucracy man, but that sounds fun too :)
09:48
<annevk>
also, another day, still not published
09:49
<annevk>
Mike is still flying around?
09:49
<jgraham>
That's superman
09:50
<jgraham>
Although I guess Mike might hve an alter-ego
09:50
<jgraham>
Would explain a lot
09:51
<zcorpan>
xml-stylesheet will be published 28th
09:52
<annevk>
as REC?
09:52
<zcorpan>
yeah
09:52
<zcorpan>
my first REC :)
09:52
<annevk>
with how many conforming implementations?
09:53
<zcorpan>
dunno
09:53
<annevk>
congrats anyway
09:53
<annevk>
I don't have any RECs
09:56
<zcorpan>
hmm validator.nu still doesn't validate xml-stylesheet
09:59
<zcorpan>
hsivonen: ^
10:07
<zcorpan>
seems like all tests in http://simon.html5.org/test/xml/xml-stylesheet/parsing/ give bogus errors in validator.nu
10:14
<hsivonen>
zcorpan: hmm. I wonder if MikeSmith landed the code...
10:14
<hsivonen>
zcorpan: the code is there
10:17
<hsivonen>
zcorpan: looks like the XML parser is doing something very bogus
10:19
<hsivonen>
zcorpan: I'm sorry. this XML parser is very, very sad
10:20
<hsivonen>
some day, I should write my own XML parser, with enough 5
10:21
<zcorpan>
did you switch xml parser at some point? it worked when mike was implementing xml-stylesheet
10:21
<hsivonen>
zcorpan: no, I think I haven't changed anything about the XML parser
10:22
<zcorpan>
ok. weird
10:24
<hsivonen>
whoa. "unless you cannot accept LGPL in your environment"
10:25
<hsivonen>
why on earth do the subclipse developers bother to maintain installation options for people who aren't OK with LGPL?
10:28
<hsivonen>
whoa! I just saw a signed Eclipse plug-in for the first time. Seemed to be a self-signed cert, though.
10:30
<hsivonen>
zcorpan: I haven't touched the XML parser since April 2009
10:38
<hsivonen>
zcorpan: are you sure you've tested without the XML declaration before?
10:38
<hsivonen>
zcorpan: the parser seems to be bogus
10:39
<hsivonen>
zcorpan: if the first four bytes are <?xm, it expects it to be the start of an XML declaration and it expect the XML declaration to be within the first 64 bytes
10:39
<hsivonen>
both expectations are thoroughly bogus, of course
10:41
<zcorpan>
hsivonen: i'm pretty sure i tested without xml decl. but maybe mike's setup injected an xml decl or something
11:02
<annevk>
http://arewefastyet.com/ interesting
11:02
<annevk>
I wonder how Opera compares
11:03
<jgraham>
annevk: That seems quite simple to find out :)
11:04
jgraham
wonders how numbers inside the browser compare to standalone shell numbers
11:18
<annevk>
Hixie, http://tantek.com/2010/294/t7/html5-whatwg-stopped-twitter-required-oauth
11:20
<benschwarz>
Hixie: P.Irish asked me to log here. Hello :)
11:33
<zcorpan>
hsivonen: if i add an xml decl, i still don't get the xml-stylesheet pi validated (no errors for a bogus pi)
11:34
<zcorpan>
http://validator.nu/?doc=data%3Aapplication%2Fxml%2C%253C%253Fxml%2520version%253D%271.0%27%253F%253E%250D%250A%253C%253Fxml-stylesheet%2520xyz%253F%253E%250D%250A%253Chtml%2520xmlns%253D%27http%253A%252F%252Fwww.w3.org%252F1999%252Fxhtml%27%253E%250D%250A%253Chead%253E%250D%250A%253Ctitle%253E%253C%252Ftitle%253E%250D%250A%253C%252Fhead%253E%250D%250A%253Cbody%253E%250D%250A%253Cp%253E%253C%252Fp%253E%250D%250A%253C%252
11:34
<zcorpan>
252Fhtml%253E
11:41
<hsivonen>
wiki.whatwg.org times out for me
11:41
<hsivonen>
zcorpan: I think the data: URL you pasted isn't complete
11:42
<hsivonen>
Hixie: (Can't contact the database server: Lost connection to MySQL server at 'reading initial communication packet', system error: 110 (database.wiki.whatwg.org))
11:42
<zcorpan>
yeah, it got split to two messages
11:58
<annevk>
I wonder if there exists more of this stuff: http://www.aminutewithbrendan.com/pages/20101018
12:02
<jgraham>
What stuff?
12:03
<annevk>
relatively short podcasts on browser stuff
12:03
<espadrine>
aminutewithbrendan is full of it
12:03
<espadrine>
(although it is javascript-oriented)
12:03
<annevk>
sure, I'm subscribed to that one and listen to most of it
12:05
<espadrine>
crockford has a series of talks going on, but we can't call them short
12:25
<jgraham>
So listening to those has given me an appreciation for one thing; if you set the volume on one <audio> element all other unplayed elements should start at the same volume by default. Or something
12:26
<jgraham>
A UI issue of course
12:34
<annevk>
and not necessarily true
12:34
<annevk>
as it depends on the stream
12:36
Philip`
would appreciate his browser not automaticall downloading many megabytes of data when loading the front page of that site, and instead waiting until he clicked 'play' on each <audio>
12:37
<Philip`>
s//y/
12:38
<espadrine>
The volume of one stream should not modify that of other streams on the same page, but it would be quite useful to have the browser remember the volumes of each stream
12:38
<espadrine>
when we go back to that page later on.
12:39
<espadrine>
I mean, it could be localStorage-ed, but that's something we inherently want for all media elements.
12:49
<hsivonen>
this XML parser is full of sadness
12:50
<jgraham>
The problem is I start one, listen to it (at reduced volume) start another and have to reduce the volume again
12:50
<jgraham>
Changing my system volume affects other applications
12:50
<jgraham>
I could probably futz about with per-application volume controls but its a lot of complexity
12:51
<jgraham>
and, on the web, the intrinsic loudness of different sound clips seems highly likely to change between pages
12:52
<jgraham>
due to different recording and so on
12:52
<jgraham>
But might be more constant on a single page
12:54
<hsivonen>
this total bogosity of the XML parser with <?xml-stylesheet hasn't come to my attention before
12:54
<hsivonen>
I guess most people use the XML declaration
13:00
<Philip`>
jgraham: Seems like the OS's job to make per-application volume controls simple
13:00
<Philip`>
(I think Vista does that already)
13:01
<jgraham>
Philip`: In the case of the web it's not clear that will work, as I said
13:01
<jgraham>
In any case I probably want per-tab volume controls
13:01
<jgraham>
if we are doing it that way
13:13
<hsivonen>
zcorpan: http://html5.validator.nu/?doc=http%3A%2F%2Fsimon.html5.org%2Ftest%2Fxml%2Fxml-stylesheet%2Fparsing%2F001.xml
13:13
<hsivonen>
zcorpan: I hope the parser sucks less now when viewed as a black box. The source code got uglier.
13:13
<hsivonen>
It seems to me that when a piece of software is bad enough, patching it makes the code look even worse
13:14
<hsivonen>
Ælfred2 is so, so sad. Too bad it's not easy to notice on surface before it's too late.
13:34
<zcorpan>
hsivonen: cool, thanks
13:36
<hsivonen>
zcorpan: thanks for reporting
13:37
<zcorpan>
hsivonen: now the Show Source feature is broken instead :)
13:37
<hsivonen>
aargh
13:42
<jgraham>
Dear lazyirc. I have a file with some lines in it. I want to provide each line of the file as the command line argument to a program. Is there some clever way to do that in the (bash) shell?
13:42
<jgraham>
So if the file looks like
13:42
<jgraham>
A
13:42
<jgraham>
B
13:42
<jgraham>
C
13:42
<jgraham>
I want to run
13:42
<jgraham>
foo A B C
13:43
<jgraham>
where foo is the executable
13:46
<jgraham>
Oh, maybe cat and backticks is enough if there are no special characters
13:47
<jgraham>
And xargs is needed otherwise
13:52
<Workshiva>
Backticks make me sad
13:52
<Workshiva>
$() all the way
14:03
<jgraham>
The shell makes me sad
14:05
<annevk>
Python
14:29
<annevk>
Pro tip: when disabling Bluetooth, do not disable wireless instead
14:29
<Philip`>
perl -ne'`program "$_"`' lines.txt
14:32
<Philip`>
Oh, I misread the question
14:34
<Philip`>
In that case I suppose it's more like perl -e'chomp(@_=<>); system("program", @_)' lines.txt which will incidentally handle special characters in the lines
14:35
<Philip`>
but that code doesn't use enough quote characters :-(
14:35
<Philip`>
(Not enough to be pleasantly confusing, at least)
16:49
<MikeSmith>
hsivonen: about the issue that Simon mentioned, I am not injecting an xml declaration or anything like that
16:50
<MikeSmith>
but I can take a look later and see if i might be able to figure out what's wrong
16:50
<MikeSmith>
maybe I am otherwise just doing something stupid in that code
16:50
<MikeSmith>
wrong assumption or something
16:50
<hsivonen>
MikeSmith: your code isn't wrong
16:50
<hsivonen>
MikeSmith: the XML parser is wrong
16:51
<MikeSmith>
ok
16:51
<MikeSmith>
oh
16:51
<MikeSmith>
well, that's not good
16:51
<hsivonen>
MikeSmith: are you sure you and Simon didn't test a W3C instance that runs the perl preprocessor on the input?
16:51
<MikeSmith>
no
16:51
<MikeSmith>
definitely not
16:52
<hsivonen>
well, I don't have an explanation why you haven't seen this before
16:52
<MikeSmith>
whatever code I had running, it should still be running at http://www.w3.org/html/check
16:52
<hsivonen>
but the XML parser is definitely broken and has been unmodified for over a year
16:52
<MikeSmith>
I see
16:52
<hsivonen>
MikeSmith: your instance shows the same bug
16:53
<hsivonen>
MikeSmith: http://www.w3.org/html/check?doc=http%3A%2F%2Fsimon.html5.org%2Ftest%2Fxml%2Fxml-stylesheet%2Fparsing%2F002.xml
16:53
<MikeSmith>
well, that's weird
16:53
<hsivonen>
I'll have dinner and then I'll try to fix the XML parser again. more correctly this time hopefully
16:53
<MikeSmith>
actually I was not running Simon's test from that side
16:53
<MikeSmith>
I was using direct input
16:53
<hsivonen>
though I'm not planning on fixing it totally correctly
16:53
<MikeSmith>
ok
16:53
<hsivonen>
it's too broken for that
16:54
<hsivonen>
MikeSmith: direct input doesn't sniff the bytes for encoding, so a different code path applies
16:54
<MikeSmith>
ah
17:11
<annevk>
complained about lack of response http://lists.w3.org/Archives/Public/public-pfwg-comments/2010OctDec/0007.html
17:11
<annevk>
apparently my last email was end of August
17:25
<cryzed>
http://code.google.com/p/html5lib/issues/detail?id=141 -> Could you guys take care of this? Or are you planning on dropping BeautifulSoup support altogether?
17:33
hsivonen
removes UTF-32 support from the XML side of Validator.nu
17:33
<annevk>
yay
17:34
<annevk>
the fight encoding proliferation front salutes you
17:35
<hsivonen>
annevk: did you notice the Gecko UTF-32 removal bug?
17:37
<annevk>
yup
17:48
<hsivonen>
zcorpan: http://html5.validator.nu/?doc=http%3A%2F%2Fsimon.html5.org%2Ftest%2Fxml%2Fxml-stylesheet%2Fparsing%2F002.xml&showsource=yes
17:48
<hsivonen>
now fixed for real, I hope
17:52
<annevk>
geez
17:52
<annevk>
gmail is getting so slow
17:52
<annevk>
it's annoying
17:52
<annevk>
maybe webapps are not it after all
17:53
<annevk>
constantly "still working" when I try to get through my labeled email
18:18
<MikeSmith>
annevk: hsivonen using Zimbra
18:20
<MikeSmith>
saucy is sähäkkä http://en.wiktionary.org/wiki/s%C3%A4h%C3%A4kk%C3%A4#Finnish
18:20
<MikeSmith>
comparative sähäkämpi, superlative sähäkin or sähäköin
18:47
<hsivonen>
abarth: b.m.o got another duplicate of http://www.w3.org/Bugs/Public/show_bug.cgi?id=10427
18:48
<hsivonen>
abarth: https://bugzilla.mozilla.org/show_bug.cgi?id=605373
18:48
<hsivonen>
has WebKit gotten duplicates of this bug?
18:48
<hsivonen>
abarth: it might be worthwhile to check what exactly the old WebKit tree builder did
18:49
<abarth>
i don't think we've gotten reports of this
18:49
<abarth>
but the page is obviously broken in webkit now too
18:56
<annevk>
http://twitter.com/devongovett/status/28426814227 -- isn't it layout that's slow?
18:57
<abarth>
dromeo tests dom speed
19:00
<hsivonen>
testing layout speed is *hard*
19:00
<hsivonen>
it's probably a bad idea to publish a layout speed benchmark if bz hasn't reviewed it
19:10
<AryehGregor>
So as of yesterday, looks like Firefox beats Chrome on Sunspider: http://arewefastyet.com/
19:10
<AryehGregor>
Very slightly.
19:12
<AryehGregor>
Awesome. I just wish we had some more benchmarks up there, and Opera and IE9 were on them too.
20:21
<jgraham>
People who say that javascript isn't slow are just making the wrong type of application. For a javascript game or audio analysis or image progessing or something, javascript speed is still a limiting factor
20:21
<jgraham>
Of course for other types of application DOM is the limiting factor
20:22
<jgraham>
But it is often pure scripting performance that limits the kind of applications that people are only just starting to write in javascript
20:22
<AryehGregor>
How does JS compare to something like Java these days on standard useless programming benchmarks?
20:24
<aho>
http://shootout.alioth.debian.org/u64/benchmark.php?test=all&lang=v8&lang2=java
20:24
<aho>
regex is awesome with v8 :>
20:24
<aho>
other than that java is typically 3-8 times faster
20:24
<aho>
or even over 50 times faster if you do something js is very bad at :>
20:25
<AryehGregor>
Hmm, so probably still room for improvement.
20:25
<AryehGregor>
pidigits is horrible for v8.
20:25
<aho>
garbage collection is probably one of those areas with some room for improvements
20:26
<aho>
java's gc is very advanced
20:26
<AryehGregor>
Interesting that memory is about the some.
20:26
<AryehGregor>
same.
20:26
<jgraham>
Possibly the implementation of pidigits for the v8 case is very bad
20:26
<jgraham>
(possibly not of course)
20:26
<aho>
http://arewefastyet.com/ <- seen that?
20:27
<aho>
jm is now faster than v8 in sunspider
20:27
<jgraham>
Needless to say these are not a good way to judge anything
20:27
<AryehGregor>
jgraham, they're a decent heuristic.
20:28
<AryehGregor>
Interesting. It looks like Java doesn't perform much worse than C in time taken, at least for these programs -- median 2x slower -- but it's murdered on memory use.
20:28
<jgraham>
AryehGregor: Well they probably have the advantage that language authors are unlikely to optimise for these benchmarks specifically
20:28
<AryehGregor>
Yes.
20:29
<jgraham>
But I think they are all very small problems compared to typical real applications
20:30
<Philip`>
Performance in real applications is typically limited by bottlenecks, which are very small problems
20:30
<AryehGregor>
Why don't they include PHP? The site is even written in PHP.
20:30
<AryehGregor>
Also would be cool to see different compilers compared. ICC really murders GCC.
20:30
<aho>
there is php on another machine
20:30
<aho>
http://shootout.alioth.debian.org/u32/benchmark.php?test=all&lang=php&lang2=v8
20:31
<jgraham>
Philip`: I don't know if that is really true. Sometimes you get quite flat performance profiles where you spend an equal amount of time everywhere
20:32
<jgraham>
I mean if you are doing n-body simulations or something then sure it is all inner-loop dominated
20:32
<aho>
the php version of reverse-complement is sorta silly... it does everything in native functions. there are like 4 functions chained together with do everything :>
20:32
<aho>
*which
20:33
<AryehGregor>
aho, well, that's sensible, isn't it?
20:34
<aho>
well, it's sorta like having solveThisSpecificProblem function
20:34
<aho>
you could say the same about the regex stuff though
20:34
<aho>
.>
20:35
<AryehGregor>
It's fair to give languages credit for having good general-purpose tools like strtr and strrev and chunk_split.
20:35
<aho>
just wanted to point out why php does surprisingly well in that particular benchmark .)
20:35
<AryehGregor>
PHP has a mediocre standard library, but it's written entirely in C, so it deserves benchmark credit for that.
20:36
<jgraham>
Hmm, I wonder if the problem with pidigits is the lack of a sensible long integer type in javascript
20:37
<aho>
yea, emulating bigint stuff with doubles is sorta clumsy
20:37
<aho>
but there is no way around that currently
20:38
<aho>
well, there are ints in most js engines under the hood but there is some drag involved
20:38
<aho>
bitops for example are always sorta slow
20:39
<aho>
(i still use xorshift as prng though) :>
20:39
<jgraham>
I think that is not true anymore since I think they are tested in sunspider :)
20:40
<jgraham>
I would expect a loop containing bitops to be JITed to native instructions
20:40
<aho>
jm was always pretty good at doing brainless math in a loop
20:41
<aho>
it's pretty much the best case for tracing
20:42
<aho>
eg 6 months ago mandelbrot stuff was about 3 times faster in jm than v8
20:44
<aho>
either way, i love the js engine wars :)
20:44
<aho>
(thanks google) :>
20:50
<AryehGregor>
Thanks Mozilla too.
20:50
<AryehGregor>
arewefastyet is great.
21:07
AryehGregor
sees this in an EULA: "You may not, in whole or in part: copy, photocopy, reproduce, translate, reverse engineer (with the exception of specific circumstances where such act is permitted by law), derive source code, modify, disassemble, decompile, or create derivative works based on the Program; . . ."
21:07
<AryehGregor>
Does that mean that the EULA is admitting that it's not prohibiting reverse-engineering except where it's already illegal anyway?
21:08
<AryehGregor>
Hmm, maybe it only means "except in cases where the law doesn't let you sign away your right to reverse-engineer". Or maybe it just means "we can sue you for breach of contract too if you do it".
21:08
<AryehGregor>
Meh, who cares.
21:13
<espadrine>
Arguably, it means that you are allowed to do so in any country that doesn't specify that reverse engineering is illegal.
21:13
<espadrine>
What EULA is it?
21:15
<AryehGregor>
Empire: Total War, a computer game.
21:16
<espadrine>
Well, they do unusual EULA then.
21:16
<AryehGregor>
Do they? How so?
21:17
<AryehGregor>
http://www.google.com/intl/en/jobs/joininggoogle/resume/index.html <-- HTML not an acceptable format? Lame.
21:17
<espadrine>
Although, iirc, Sun had a clause that forbid nuclear facilities from using their stuff.
21:18
<Ms2ger>
Opera does
21:18
<Ms2ger>
Also, flight control, IIRC
21:18
<AryehGregor>
Fascinating.
21:18
<jgraham>
Opera also accepts resumes in HTML form
21:18
<jgraham>
just sayin'
21:19
<AryehGregor>
I'd just have to move to Scandinavia, is that all?
21:19
<Ms2ger>
It would be pretty sad if they didn't
21:19
<AryehGregor>
Probably Google would too, but no sense in not following instructions.
21:20
<jgraham>
Well I say "accepts" though for most technical jobs I mean "requires"
21:20
<AryehGregor>
See, now that makes sense.
21:20
<jgraham>
AryehGregor: Well not necessarily. But most jobs are in Scandinavia
21:20
<AryehGregor>
And have someone manually inspect the source code and reject the application if they use <FONT COLOR="RED">.
21:20
<AryehGregor>
None in, say, New York or Israel?
21:21
<jgraham>
AryehGregor: http://www.opera.com/company/jobs/
21:21
<AryehGregor>
So that's a no, then.
21:22
<jgraham>
And you can infer what you like about the <font color> stuff from http://www.opera.com/company/jobs/opening/67/
21:22
<AryehGregor>
"Your CV and covering letter should be in HTML5/CSS. That is your first test."
21:22
<jgraham>
"Your CV and covering letter should be in HTML5/CSS. That is your first test."
21:22
<AryehGregor>
Haha.
21:22
<AryehGregor>
Awesome.
21:23
<AryehGregor>
I also bet you get points off if you do -webkit-border-radius: 1em; -moz-border-radius: 1em
21:23
<jgraham>
Well I'm not involved with recruitment
21:24
<jgraham>
But I guess one would think twice before sending in something "Best Viewed with Google Chrome"
21:37
<AryehGregor>
I just realized I have no idea how to write a decent resume.
21:40
<jgraham>
You start with an indecent one and add clothes
21:45
<aho>
:>
21:48
<Philip`>
AryehGregor: At least in the UK, there are certain cases where copyright law explicitly allows you to reverse-engineer programs (e.g. for interoperability) regardless of any other agreements that attempt to prohibit you from doing that
21:49
<Philip`>
so I guess the agreements want to be clear that their terms should not be completely ignored, since they're still prohibiting all the other cases that aren't those certain cases
21:50
<Philip`>
Oh, actually that's about "decompilation" rather than reverse-engineering
21:50
<Philip`>
but close enough
21:51
<gsnedders>
MikeSmith: I have some vague idea about pronounciation, but that's the extent of my knowledge of Gaelic
21:52
<MikeSmith>
ok
21:53
gsnedders
has some vague suspicion his CV was HTML 4.01 Strict
21:53
<gsnedders>
(But I can't remember)
21:55
<jgraham>
AryehGregor: (If you wanted more useful CV advice possibly 'nobody cares about what you did at high school' and 'imagine you have to read 100 of them and want to find all the interesting stuff quickly' would be better)
21:55
<AryehGregor>
Noted.
21:57
<AryehGregor>
I guess my CV should clearly say "Okay, I'm just out of college, but I actually have lots of experience at this stuff as a volunteer."
21:57
<MikeSmith>
AryehGregor: I think you should submit your CV as an animated movie created using xtranormal
21:57
AryehGregor
makes sure to use goo.gl for his URL shortening, and makes a mental note to change that if he submits to anyone other than Google
21:58
<gsnedders>
AryehGregor: I'd say more, look at this experience, and hey, I went to college!
21:58
<AryehGregor>
MikeSmith, I think that kind of resume is probably better suited to people applying for jobs as designers.
21:58
<jgraham>
gsnedders: More like "hey I actually finished college" :p
21:58
<gsnedders>
AryehGregor: Having gone to college is unlikely to get you a job. Having experience is probably more important.
21:59
<gsnedders>
jgraham: Hey, I'm still at university, and went to all my lectures today!
21:59
<gsnedders>
(Turns out programming lectures are still ungodly boring)
21:59
<jgraham>
(that isn't true I think)
21:59
<jgraham>
(not about gsnedders lectures)
21:59
<jgraham>
(but lots of people get jobs with degrees and little experience)
21:59
<AryehGregor>
gsnedders, well, I don't think anyone is going to be fooled into thinking I have lots of professional experience at this point.
22:00
<jgraham>
(having a degree *and* experience is obviously better)
22:00
<AryehGregor>
I mean, I should come off as being much better than some average dude just out of college.
22:00
<AryehGregor>
But probably not as good as someone with five years' of actual job experience.
22:00
<jgraham>
(having just experience depends a lot on who you are applying to)
22:00
<AryehGregor>
(except that if I'm applying specifically for web specs, probably like twelve people have at least five years' of job experience with that, so)
22:01
<AryehGregor>
(half of them are probably in this room)
22:01
<gsnedders>
Yeah, sure. And my point about mentioning the experience first is more that you want to set yourself apart. Whoever is looking at it probably has 100 others to look at, all starting with, "I went to the university and got put in a box".
22:01
<gsnedders>
AryehGregor: To actually edit specs full time, or?
22:01
jgraham
didn't get put into a box at university
22:01
<jgraham>
although there was probably a societry for that if it's your thing
22:02
<gsnedders>
jgraham: Yes you did, you're the same as everyone else!
22:02
<gsnedders>
(There again, you'll probably claim I can't reference songs without girls around.)
22:02
<gsnedders>
(Again.)
22:02
<MikeSmith>
AryehGregor: under the Entrepreneurial Experience heading, I recommend not listing any marijuana grow operations you might be running
22:03
<AryehGregor>
MikeSmith, but if I'm doing something more unusual and enterprising, like smuggling cocaine, you think that would be a good thing to say?
22:03
<jgraham>
gsnedders: (the irony of the song is that it is almost entirely the wrong way roungd, I think)
22:03
<MikeSmith>
also, experience with making money trading smuggled contraband in country jail is often to valued quite a highly as you might think it would be
22:03
<jgraham>
(I suspect people who go to university end up leading way more diverse lifestyles than people who don't)
22:03
<gsnedders>
jgraham: (Then there's the irony of who keeps singing it :P)
22:03
<MikeSmith>
AryehGregor: just use code words
22:04
<AryehGregor>
jgraham, it probably depends on how you measure diversity.
22:04
<jgraham>
MikeSmith: Like when you say "Man mangement skills" and you mean "I'm a sociopath that will make CEO one day"
22:04
<MikeSmith>
heh
22:05
<AryehGregor>
MikeSmith, so do you think I should be aiming to make it into thedailywtf.com's next round-up of resume submissions, or what?
22:06
<jgraham>
AryehGregor: Sure, it's all about getting your name known in the right circles
22:07
AryehGregor
realizes that goo.gl provides public click profile data, so if he includes a goo.gl URL in a resume that he only submits to Google, he'll have a good idea of how many people read it and where they are.
22:07
<AryehGregor>
The irony is delicious.
22:07
<AryehGregor>
jgraham, sadly, TDWTF tends not to include names.
22:08
<gsnedders>
AryehGregor: Seriously though, what specifically are you trying to get a job doing?
22:08
<MikeSmith>
AryehGregor: there's no such thing as bad publicity
22:08
<Ms2ger>
You could just have your resume on your website and send yourself an email for every hit
22:08
<Ms2ger>
(Like Hixie)
22:08
<AryehGregor>
gsnedders, my general goal at this point is to get a job where people will pay me money. Possibly this will involve me performing some activity they desire in exchange, preferably one that I find at least moderately enjoyable.
22:09
<AryehGregor>
This seems like a fairly conventional course of action.
22:09
<Philip`>
Ms2ger: If you do that, you should be very careful not to mention it in the presence of people who might then find it amusing to set up a script that downloads the page and loops several times a second
22:10
<AryehGregor>
Realistically, something in web standards seems like a good bet, because there's hardly a lot of competition. Failing that, programming is fine too.
22:10
<jgraham>
So far it is not clear that you have excluded, say "high class male escort" from your list of possibilities
22:10
<AryehGregor>
Well, I doubt that would be very competitive pay, given my skill set.
22:10
<AryehGregor>
Also given my incredibly nerdy appearance.
22:12
<gsnedders>
Hey, I'm sure some are into that.
22:13
<jgraham>
Maybe not enough for a regular income
22:16
<Hixie>
hm, did anyone write a counter-proposal for the rel registry?
22:16
<Hixie>
we're a few days past the deadline
22:18
<othermaciej>
TabAtkins said he was going to, but I don't know if he did anything yet
22:19
<hober>
at the same time, the wg doesn't seem enthusiastic about the IETF registry
22:28
<Hixie>
well, i sent a null CP
23:56
<annevk>
AryehGregor, we hired someone who works from New York
23:57
<annevk>
AryehGregor, it depends on the position really, but we have several people working from home, including people working on the Core engine
23:57
annevk
works from home too
23:57
<annevk>
mostly
23:59
gsnedders
works from home