[07:31] <pwnguin> bryce: i have a very random question for you. which do you think is faster to render? svg or png?
[07:42] <bryce> depends on the complexity of the svg
[07:43] <bryce> pngs have some overhead due to having to decompress them, but aside from that the resultant raster is relatively fast to blit
[07:44] <bryce> svg supports advanced functionality like filters, gradients, etc. which can be processor intensive for some chipsets
[07:44] <bryce> however a really trivial one - just some simple shapes of solid colors - could actually render faster
[07:44] <pwnguin> lets say the eclipse icon level
[07:45] <pwnguin> a couple circles, some lines and a highlight
[07:45] <bryce> it might be faster - again though, it depends on a lot of factors so it's best to just try it and measure
[07:46] <bryce> however for icons, they're so tiny it's usually a moot point one way or the other
[07:46] <pwnguin> im thinking about making a push to get rid of more svg icons
[07:46] <pwnguin> err
[07:46] <pwnguin> its kinda late.
[07:47] <pwnguin> get rid of png / xpm for svg
[07:47] <pwnguin> im just trying to look at the angles
[07:47] <pwnguin> there's complaints that the ubuntu logo is too complex for the screensaver
[07:48] <bryce> for inkscape, an advantage we saw is that we could put all the icons into a single .svg file, render that whole file in one go, and then slice out the rasters for the icons
[07:48] <pwnguin> "slice out"
[07:48] <pwnguin> ah, render high then drop lines
[07:48] <bryce> some applications have each icon a separate .png file, so turning all that into a single file I/O operation can be quite significant
[07:49] <bryce> filesystem I/O tends to be a major performance killer, so any optimization which allows you to avoid file I/O at the expense of processor or memory, tends to work out pretty well
[07:49] <pwnguin> so svgz
[07:50] <bryce> in theory you could do the same thing with .png's and put them all in the same file and then use offsets to slice out the individual images
[07:50] <bryce> I know a lot of games do that for instance
[07:51] <pwnguin> i need to get nussbaum's cluster
[07:51] <bryce> an advantage with svg is that editing is a lot easier, since you don't have to worry so much about pixel placement, and can really easily scale, copy and paste bits from other icons (file folders, letters, etc.)
[07:51] <pwnguin> and do an archive query on the size of svg icons
[07:52] <bryce> of course, raster icons do have their advantages, like if you want to do hinting and optimize for specific sizes/resolutions
[07:52] <bryce> to me the best argument favoring using svg for icons is that it makes it sooo accessible for artists to contribute
[07:52] <pwnguin> that's my next question: does anyone actually optimize icons?
[07:53] <bryce> some time look through inkscape's menus and you'll see there's hardly any menu items that *don't* have icons
[07:53] <bryce> yeah some people do
[07:54] <pwnguin> well, how many packages is the relevant question
[07:54] <bryce> I'd guess all the major ones.  For joe random open source project, they probably don't.
[07:56] <pwnguin> i mean, i dont really care about the icons in a program
[07:57] <pwnguin> but the .desktop / launcher icons are branding
[07:58] <pwnguin> hmm
[07:59] <pwnguin> maybe its not nessecary to change out the icons, but just make sure they all have a scalable one available
[08:03] <pwnguin> bryce: well thanks for the input. certainly plenty of detail to go on
[08:57] <bryce> sure, I'd love to hear how things develop
[08:58] <bryce> pwnguin: if you're really interested in svg icons, make sure to hang on #inkscape... lotta svg icon fanatics there with lots and lots of advice.  tell folks you work with me on ubuntu-x and they should treat you well :-)
[09:10] <pwnguin> i dont do all that much work
[09:10] <pwnguin> more like sophisticated stalking
[09:11] <bryce> it's appreciated nonetheless :-)
[09:11] <pwnguin> im thinking of applying for ubuntu membership
[09:19] <bryce> pwnguin: cool!  :-)  Yeah I think you'd have no trouble
[09:20] <pwnguin> but i'd need a few people to show up to declare me totally awesome and a super rad dude
[09:20] <bryce> pwnguin: count me in, just let me know when
[09:20] <pwnguin> i have no idea
[09:20] <pwnguin> the next meeting is "undecided"
[09:20]  * bryce nods
[09:21] <bryce> I think they go once a month, and the last meeting was a week or two ago
[09:21] <pwnguin> yea i noticed
[09:21] <pwnguin> a ton of loco people on the planet
[09:22] <bryce> in any case, I'll be happy to support you.  If you find out when the meeting is, let me know.
[09:23] <bryce> pwnguin: do you have plans to go for MOTU?
[09:23] <pwnguin> sure
[09:23] <pwnguin> MOTU is kinda wierd
[09:23] <pwnguin> i dont have wide interests
[09:24] <pwnguin> everyone says they're more generalists
[09:24] <pwnguin> plus, im fine with having my work peer reviewed
[09:25]  * bryce nods
[09:25] <pwnguin> i do have a ppa
[09:25] <pwnguin> i guess the way to put it is a square
[09:25] <bryce> I did Member -> MOTU -> Core Dev, so for me it was a good stepping stone.
[09:26] <pwnguin> stack ubuntu on the bottom, debian in the middle and upstream at the top
[09:26] <bryce> it let me get some experience uploading to universe and doing sponsorships, and while I didn't do a ton of MOTU work, it was good preparation
[09:27] <pwnguin> most of ubuntu is horizontal; I'm more interested in a vertical approach
[09:28] <pwnguin> my fingerprint reader for example, I follow the ML and the debian team's work
[09:28]  * bryce nods
[09:28] <pwnguin> i pulled from debian experimental and some patches on the ML / bugzilla
[09:29] <pwnguin> thats all been superceded though since thinkfinger was brought into main
[09:36] <superm1> well and it will be changing for intrepid pwnguin.  fprint is on its way :)
[21:08] <pwnguin> superm1: i know
[21:09] <pwnguin> superm1: does anyone actually discuss these things, or do they just do it and hope people notice?
[23:32] <superm1> pwnguin, i talked to Keybuk about it