00:02
<Hixie>
it's a complicated set of scripts
00:02
<Hixie>
split across multiple machines
00:02
<Hixie>
why?
00:14
<zcorpan>
i have an image that got absolutely ruined by quality 0 jpeg with gimp
00:25
<WeirdAl>
Hixie - well, I'd like to be able to locally break that 6MB file up into chapters
00:25
<WeirdAl>
and I'd rather use official scripts than try to homebrew it
00:25
<WeirdAl>
Python would be really good at it, I'd think...
00:26
<WeirdAl>
if the scripts were checked in, that'd be good enough for me to work with (I think)
00:29
<WeirdAl>
hmm, I'll have to file a bug on XHTML serialization - it doesn't specify what should happen with the XML prolog, if there ever was one
00:30
<WeirdAl>
when parsing the DOM, the <?xml ?> prolog is usually lost.
00:31
<Youdaman>
g'day
00:32
<Youdaman>
divya: thank you for your informative video via MarakanaTechTV :)
01:04
<zcorpan>
so from my testing, with the image i played around with, using quality 0, 10, 20, results in quite blocky artifacts which look horrible. however, quality 40 is OK and gives a better image zoomed in compared to smaller quality 90 image.
01:07
<zcorpan>
so using a 2x image with 40 quality results in similar-filesize with better quality for retina (or zoomed in), while using 4x with 10 quality results in similar-filesize with horrible quality everywhere
01:07
<zcorpan>
(compared to 1x 90 quality)
01:08
<zcorpan>
(all 4:2:2 subsampling and unchecked most checkboxes in gimp's export jpeg)
01:09
<zcorpan>
(using 4:2:0 subsampling seems to reduce file size but doesn't degrade quality noticeably)
01:42
<zcorpan>
http://simon.html5.org/dump/compressive/test.html
01:42
<Hixie>
hm, weirdal left
01:43
<Hixie>
size 25% isn't loading properly for me
01:43
<Hixie>
quality 10 is a disaster on the wall
01:45
<zcorpan>
problems uploading...
01:47
<zcorpan>
sigh. i'll try again tomorrow.
03:13
<GPHemsley>
apparently filesize equates to "weight" now?
03:17
<GPHemsley>
wait, so, WeirdAl doesn't want to reverse-engineer the scripts powering the spec that was written because people were always having to reverse-engineer how scripts were processed?
03:34
<SamB_MacG5>
"complicated set of scripts split across multiple machines" sounds like a bad idea ...
03:35
<SamB_MacG5>
(though I guess that's also par for debian.org/debian.net ...)
07:25
<annevk>
Hixie: I'm afraid of me bringing it up again later :p
07:25
<annevk>
(re WorkerUtils)
07:53
<divya>
Youdaman: thanks you are welcome is that why you are here :)
07:53
<divya>
if so i am glad :)
07:55
<hsivonen>
is there a backchannel for the plenary?
07:57
<annevk>
#tpac-chat on W3C IRC
07:57
<hsivonen>
thanks
07:57
<annevk>
there's also #tpac-treehouse on Freenode, but more quiet
07:58
<Ms2ger>
Anything happening?
07:59
<hsivonen>
Ms2ger; we should listen to analysts
08:04
<Youdaman>
divya: yep that's why i'm here :)
08:06
<smaug____>
hsivonen: and there is also #tpac
08:07
<annevk>
modular modular modular
08:07
<annevk>
because if you split a coherent doc in two, it's modular...
08:08
<hsivonen>
#tpac is for minuting
08:10
<annevk>
teehee, I think I can merge parse and interpret soonish
08:11
<divya>
Youdaman: naice!
08:11
<divya>
any new memes yet?
08:11
<divya>
i see no source for memes tho.
08:12
<hober>
annevk divya: i guess we could make some kind of whaambulance meme
08:12
<divya>
ahah what does it mean hober
08:12
<annevk>
http://lists.w3.org/Archives/Public/public-script-coord/2012OctDec/0110.html
08:13
<annevk>
divya: ^^
08:14
<divya>
waaah
08:14
<divya>
naice
08:14
<divya>
WHAAMBULANCE
08:37
<nessy>
http://www.youtube.com/watch?v=IT0ofNU1NzA
08:37
<nessy>
I actually had to look that one up ;-)
12:28
<smaug____>
"Extended DRM"
12:29
<smaug____>
must be something awesome :/
12:33
<annevk>
uhuh
12:33
<annevk>
zcorpan: went okay I think, considering the TPAC
12:34
<annevk>
the topic, lol
12:34
<zcorpan>
cool
12:34
<Ms2ger>
The TuPAC?
13:10
<[tm]>
wtf "ERROR Closing Link: 91.217.168.220 (Too many user connections (global))"
13:10
<[tm]>
from irc.freenode.net
13:22
<annevk>
yeah someone should email the Freenode staff with our IP
13:22
<annevk>
to ask for an upgrade
13:22
<annevk>
had the same problem yesterday
13:22
<[tm]>
annevk: no way to get around it from the client side, right?
13:22
<Ms2ger>
ssh, I guess
13:24
<[tm]>
Ms2ger: yeah that's what I'm doing here
13:27
<smaug____>
drm is just pure evil
13:29
<tantek>
not sure if drm is pure evil, or pure stupidity.
13:30
<hober>
that or isn't an xor :)
13:44
<annevk>
[tm]: no, we had the same problem at Opera, I emailed them telling them we're Opera and we like their service and would need upwards of 200 allowed connections
13:44
<annevk>
[tm]: that's about all it takes
13:45
<gsnedders>
Can you not just msg some online staff?
13:46
<annevk>
well I couldn't connect at the time :)
15:09
<GPHemsley>
If anyone gets a chance in between arguments, feedback welcome on my latest commit to mimesniff.
15:14
<annevk>
GPHemsley: hard to review :/
15:15
<GPHemsley>
yeah, the stupid differ b0rked
15:15
<GPHemsley>
marked sections as changed that hadn't been changed
15:15
<GPHemsley>
no idea why
15:15
<annevk>
a bunch of things like "a video type supported by the user agent" are not linked
15:16
<annevk>
GPHemsley: whitespace?
15:16
<GPHemsley>
the majority of the changes are in the "Unknown media type" section
15:16
<GPHemsley>
which I completely rewrote
15:16
<GPHemsley>
those unlinked terms were like that before this change
15:17
<GPHemsley>
I haven't defined them yet
15:17
<annevk>
k
15:17
<GPHemsley>
all of 5.2 needs review
15:17
<annevk>
in general I guess the cleanup looks good, but it might take a while before you find someone to review that
15:17
<annevk>
maybe once people write tests for it
15:17
<GPHemsley>
a few definitions were added in 3
15:17
<GPHemsley>
yeah
15:18
<GPHemsley>
incidentally, where would those tests live?
15:19
<annevk>
Hixie: looks like we need quirks.spec.whatwg.org
15:19
<annevk>
Hixie: ownership -> annevankesteren again
15:20
<GPHemsley>
abarth: Do you recall if the order of the list of type patterns is significant in mimesniff?
15:20
<GPHemsley>
abarth: (There doesn't appear to be much overlap between the patterns that could result in a precedence conflict, IMO.)
16:52
<abarth>
GPHemsley: I think the order is mostly not important
16:52
<GPHemsley>
ah, good
16:53
<abarth>
GPHemsley: That's something you should check before changing, however
16:53
<GPHemsley>
ah, not so good
16:53
<GPHemsley>
:P
18:08
<annevk>
Hixie: I also want notifications or notification
18:08
<annevk>
Hixie: given that we have "Fullscreen" I think "Notification" makes sense
18:12
annevk
is looking forward to next week
18:12
annevk
wants to solve host names
18:27
<Hixie>
annevk: "fullscreens" wouldn't make sense, would it?
18:27
<Hixie>
"notifications" seems more correct to me, but i dunno, your call
18:27
<annevk>
okay
18:27
<Hixie>
notifications. and quirks. ?
18:28
<annevk>
for now, yes
18:28
<annevk>
tantek wants to add some stuff, but he's not yet ready I think
18:28
<annevk>
thanks
18:28
<annevk>
gonna go back to the reception thingie now
18:28
<Hixie>
is zcorpan giving up on quirks?
18:29
<annevk>
no, W3C gave up on him by not allowing CC0, so he wants to move it here
18:29
<Hixie>
aah, ok
18:29
<annevk>
but he wants to reuse my setup for github
18:29
<Hixie>
so your user?
18:29
<annevk>
yes
18:29
<Hixie>
roger
18:30
<Hixie>
(when was quirks ever a w3c spec?)
18:31
<Hixie>
both done
18:31
<Hixie>
let me know when to update the specs list
18:33
<Hixie>
oh, i see, they just don't want people to do drafts under a non-CLA license
18:33
<Hixie>
ok
18:33
<Hixie>
well that's a non-issue, we only need the FSA from the w3c
19:12
<GPHemsley>
Hixie: Thanks for all the bugs! :P
19:13
<GPHemsley>
Hixie: A few of them are just wording issues, so I'll need your help figuring out how to say what I mean.
19:14
<Hixie>
sure
19:35
<Hixie>
MikeSmith: https://www.w3.org/Bugs/Public/show_bug.cgi?id=18338#c2 (please feel free to close the bug or reassign it if you think it shouldn't be in html after all)
19:35
MikeSmith
looks now
19:36
<MikeSmith>
Hixie: I guess I've been waiting to hear what hsivonen thinks of that one
19:36
<MikeSmith>
I vaguely recall him saying that he thinks it'd good to have the HTML spec say something about this
19:37
<MikeSmith>
anyway I'll ask him
19:37
<MikeSmith>
gotta drop off soon
19:45
<GPHemsley>
Hixie: Wasn't sure if you wanted me to move the components for the bugs assigned to do; I didn't.
19:45
<GPHemsley>
Hixie: I've responded to all the bugs you filed.
19:46
<Hixie>
k
19:47
<GPHemsley>
Hixie: P.S. Don't forget MIMESNIFF is also referenced here: http://www.whatwg.org/specs/web-apps/current-work/multipage/iana.html
19:47
<GPHemsley>
(Not sure if you need to update that, except maybe the name of the spec.)
19:58
<Hixie>
that section is not long for this world
20:27
<Ms2ger>
Hixie, do you mind us being crazy? :)
20:35
GPHemsley
proposes having a media type 'font'.
20:36
<GPHemsley>
(Also 'archive')
20:36
<GPHemsley>
font/ttf
20:36
<GPHemsley>
archive/zip
20:36
<GPHemsley>
etc.
20:38
<Ms2ger>
Heh
20:41
<GPHemsley>
Ms2ger: Unrealistic?
20:42
GPHemsley
thinks too much gets lumped into 'application'
20:42
<Ms2ger>
I've seen people talk about font/ for years :)
20:44
<GPHemsley>
ah
20:44
<GPHemsley>
well
20:44
<GPHemsley>
hmm
20:45
<GPHemsley>
If I spec it, they will come?
20:45
<GPHemsley>
"A font type is any valid media type where type is equal to "font" or the media type portion is equal to one of the following: ..."
20:46
<GPHemsley>
Ms2ger, Hixie: What do you think?
20:47
<Ms2ger>
I care very little :)
20:48
GPHemsley
assumes that is support.
20:48
<GPHemsley>
;)
22:01
<Hixie>
GPHemsley: i agree, if we assume mime types make sense at all, that there should be a font/ top-level type, but i'm not sure the mime sniffing spec is where it should be. But then if you want to replace the IANA MIME type registry and all the MIME machinery, it would make sense to have a new spec that subsumes mime sniffing and all those things.
22:01
<Hixie>
GPHemsley: might want to approach the IETF first to see if they're amenable to considering that first, though
22:01
<Hixie>
GPHemsley: also, might be worth getting the mime sniffing stuff solid first :-P
22:34
<hsivonen>
epub has prior art for font types.
22:34
<hsivonen>
application/vnd.ms-opentype and application/font-woff
22:35
<hsivonen>
better not reinvent that stuff
22:40
<hsivonen>
I don't know where the idpf pulled those out of, but this stuff has been made up already
23:55
<zcorpan>
curious about the CLA attribution requirement: "That attribution must include, at minimum, the Specification name and version number."
23:56
<zcorpan>
so if a CLA spec doesn't have a version number, it's not possible to give it attribution
23:56
<zcorpan>
yet there is no requirement to give a spec a version number, afaict
23:57
<zcorpan>
(i guess the same argument could be made for the specification name, but that hasn't actually happened)