=== BHSPitMonkey_ is now known as BHSPitMonkey [16:45] hi! [16:53] hi [16:53] what's up thorwil [16:53] dilomo: weekend :D [16:54] party time! [16:55] btw how do you find the min/max/close buttons as style: [16:55] http://img187.imagevenue.com/img.php?image=15954_Screenshot_122_1lo.jpg [16:56] hi all [16:57] I'm wondering whether to change them or not [16:57] lucazade: hi [16:57] them? [16:57] the buttons :) [16:57] my English is not perfect [16:58] dilomo: the top of the active tab looks like there is a left-over of those colored marks. it's too broad or dark [16:58] colored marks? [16:59] what do you mean the 1px line? [16:59] dilomo: with colored marks i mean the strips above the active tab like seen in some clearlooks versions [17:00] aha so by too broad you mean ther is too much space? [17:00] dilomo: actually, the active tab page doesn't give any clear sign of being sunken in or raised [17:01] I know there is some kind of confusion in the tabs [17:01] cimi told me that ;) [17:01] dilomo: look at the title of the active notebook tab. the border above it is broader then the border on the side. it looks strange [17:02] dilomo: so do you want the active page to be sunken in or raised? [17:03] I want it sunken [17:03] I prefer raised tabs [17:03] but I can't make it [17:03] as the overlap will go away if I change some properties [17:04] such as GtkNotebook::tab-overlap [17:04] dilomo: you will need a clear shadow on the borders towards the light direction and highlights on the opposite. pretty much the reverse of what happens for buttons [17:06] thorwil: you are righ some tweaking is needed for the tabs [17:07] thorwil: but back to the question: Do you like the min/max/close buttons? [17:07] thorwil: and if not what you think should be done [17:08] dilomo: i do not. they are too busy and the differing height muddies the target areas [17:08] but don't you think that they will become too boring if they are the same height? [17:09] dilomo: i don't look at window buttons to get excited and guess few people do ;) [17:09] :D [17:10] dilomo: the focused window ... the single X button there looks fine. allthough sunken look is strange for a click target [17:11] dilomo: the single button looks rounded, bu the right side of the trio doesn't [17:11] it looks hexagonal [17:11] that is a problem indeed because I make a gradient in inkscape form black to white [17:12] and there is one place where the values of thr grad equeals the gray bg [17:12] and that roundness is lost [17:12] Don't know how to fix this [17:13] dilomo: b/w overlayed gradients rarely look as good as specifically coloured gradients [17:13] coloured? [17:14] dilomo: that is because the balance of colors varies from lit to dark areas in the real world in all common lighting situations [17:15] dilomo: i'm talking about a generic gradient that is meant to work on different backgrounds vs one that is made to fit for only one specific background [17:15] so you suggest me to make the beveal effect using some other colors than b/w [17:15] yes [17:15] good idea :) [17:15] dilomo: study shadows in the real world. they often appear blueish [17:16] yes (because the sky is blue) but in all warm theme this is not good approach [17:18] dilomo: have a look at http://www.itchstudios.com/psg/art_tut.htm [17:18] Light stuff [17:19] that's cooooooool! thanks [17:19] np :) [17:19] very nice stuff [17:19] bokkmarked [17:20] dilomo: there's a risk to end up with a grey mess when working with shadow color. but the other extreme is a sterile and synthetic look [17:21] an mine is on the sterile side ? [17:21] yep [17:22] dilomo: your light direction is straight from above, right? [17:22] yes [17:26] * thorwil -> dinner [17:27] * dilomo too [17:32] it's 6:30 PM [17:32] :-| [17:32] dinner? :\ [18:17] _MMA_: hi! i'm right now looking at the svg for Sebastian's playing cards. he managed _almost_ perfect symmetry for the heart shape [18:22] but is wasteful with nodes :) [18:22] <_MMA_> :) [18:23] <_MMA_> Make any revisions you like. I'll up them. [18:24] _MMA_: nah, i will just give him example with how few nodes it can be done. not for performance considerations, but to make his live easier [18:25] Cimi: I din't ave lunch today so it's dinner for me ;) [18:25] <_MMA_> Sure. It's always great to streamline. :) [18:26] <_MMA_> thorwil: A heart could be done with 2 nodes couldn't it? [18:26] _MMA_: not quite. makes the round parts ... not round [18:27] 4 nodes [18:27] Cimi: I'm obviously still hungry because I eat letters too :) [18:27] he has 18 [18:28] thorwil: but the icon is good [18:28] dilomo: yes. i'm always feeling a bit uneasy criticizing him, as he's better at drawing icons by now :) [18:29] thorwil: :) [18:29] but the brush is not good [18:29] <_MMA_> thorwil: Naa... I think he welcomes it. [18:29] it should look wooden [18:29] not plastic [18:30] _MMA_: i know. he also has a great attitude. only wish he would join us here [18:30] <_MMA_> yeah [18:30] <_MMA_> dilomo: I have brushes with plastic handles. :) [18:30] yes but they are not classic [18:31] I have too bit the wooden ones are gourgeous [18:31] <_MMA_> So it depends on what you're going for. :) *You* want a more classic look. He might not. :) [18:32] _MMA_: probably [18:32] I think that the biggest icons should have more details [18:33] <_MMA_> dilomo: Feel free to raise the question though. Just dont be like Cimi and call it "ugly" or something. :P [18:33] gh [18:33] <_MMA_> hehe [18:34] Cimi: _MMA_ is right you are a bit hard on newcomers [18:35] <_MMA_> dilomo: I think it's a bit of a language barrier sometimes. Needs nicer words to express dislike. :) [18:35] though sometimes I need something harder than ugly, maybe horrible? :D [18:36] :D [18:36] <_MMA_> ;) [18:36] something like "ooh after seeing that theme, my eyes refuse to stay open" :D [18:37] <_MMA_> "My eyes just vomited all over the screen." [18:38] omg [18:38] that sounds great ;) [18:38] <_MMA_> :P [18:38] i had to clean my eyes with alcohol after looking at that! [18:39] ahahah [18:39] give me a whiskey please :P [18:39] your speech gentlemen is embarrassing [18:40] ;) [18:40] what is a speech-gentleman? ;) [18:41] :) [18:44] thorwil, if you're planning about a default-theme or something similar work on mockups rather than looking on *how to implement with the css engine*. that engine will be just like the pixbuf engine but with few features more (just like experience engine): it is really important for newcomers but not for fast and default themes. [18:45] Cimi: i intend to combine both [18:45] what do you mean? [18:46] Cimi: most of the How for the css-engine is worked out already and a template for it will also be a mockup of all important widgets [18:47] what I meant is that let's keep on helping people understand how that experimental engine works, because it will be interesting for a lot of new guys [18:47] but for your ideas of a default theme let's still rely on cairo for drawing, and with custom code inside the engine [18:49] you just need to find out a good cairo coder, then you can draw a lot of images and effects through the lib [18:49] guys may i ask you how much is ugly the theme i'm making? :) http://img176.imageshack.us/img176/4453/uglyvn7.png [18:50] i know is not an original effect [18:50] just wondering [18:50] lucazade, more lightborder_ratio on the tabs [18:51] yep [18:51] not satisfied of tabs [18:51] It remind me of a theme ... [18:51] a lot of [18:51] :) [18:52] Cimi: regarding finding a good developer, i'm quite happy to work with robsta so far ;p it's not exactly easy to find someone to implement your own ideas [18:52] thorwil, I mean for the cairo drawings [18:52] anyway robsta is a great guy [18:52] lucazade: but the grey color is too dull, don't you think? [18:52] dull ... means dark? [18:53] no dull like fog [18:53] ok [18:53] mmm maybe :-/ [18:53] Cimi: i fully expect that the css-engine will be only useful for prototyping for quite a while. but that's exactly what i need [18:54] thorwil, the only things that are actually difficult to implement with cairo are shadows bigger than 2-3 pixels [18:54] I've written an algorithm, it will be in the next gnome-do (0.7.0) but it's not exactly as using a gaussian blur [18:55] and it will be quite complicated for every button [18:55] a lot of code and drawing operations :( [18:55] * Cimi wants convolution filters in cairo [18:56] * dilomo wants gradient persets in Inkscape [18:56] too bad i use blur in the button template already [18:56] * _MMA_ wants a pony!! Yay! [18:57] lol [18:57] thorwil, remove the blur :( [18:57] blur is basically a convolution with a gaussian kernel [18:58] it dramatically depends on the performance of your system [18:58] because it's a long math operation [18:58] Cimi: what i'm doing with it could be approximated with 9 areas and gradients, i guess [18:59] Cimi: radial gradients in the corners, linear one on the edges. all running out to zero alpha [18:59] * _MMA_ hopes this isn't a "We need to keep in mind systems older than 5 years." opinion from Cimi. [18:59] thorwil, I've used log for similar things [18:59] _MMA_, it's not 5 years old [18:59] Cimi: log? [18:59] it is 0 months old [18:59] thorwil, logarithm [19:01] _MMA_, you need to do an integral between your original function/image with and a gaussian exp [19:01] <_MMA_> Cimi: Sorry. That was probably too much English for you. I mean, we need to be able to code for the future. Not the past. And what is "too old" is up for debate. [19:01] it's absolutely a long calculation [19:01] _MMA_: it's quite obvious in inkscape that blur is damn expensive [19:01] that's why there's no operating system that does this in software rendering [19:02] vista uses blur [19:02] but not with the software rendering [19:02] <_MMA_> thorwil: You usin' SVN? Its 50% faster now. Multithreaded. :P [19:02] it used the shaders inside the GPU [19:02] *uses [19:03] _MMA_, it's not about programming for the future [19:03] _MMA_: that will surely help me much on this single core :) [19:03] _MMA_, convolutions MUST NOT be coded to work with software rendering [19:03] <_MMA_> thorwil: doh! :) [19:03] Cimi is right [19:03] you have probably read about DX10on9 [19:04] no [19:04] well for windows 7 they prepeare [19:04] a software layer that will [19:04] simulate SX10 on SX9 videocards [19:05] or even on machines with no GPU [19:05] <_MMA_> Cimi: Sure. Like I said, I *hoped* you weren't talking about holding code back because of the past. ;) [19:05] DX* [19:05] dilomo, blur operations can be done with ati radeon >= 95xx, geforce fx and intel x3100 [19:05] _MMA_, oh ok, I misunderstood again [19:05] sorry [19:05] they tried to do the Crysis thing on 8 core i7 [19:05] <_MMA_> :) [19:06] and guess the result ~7.9 fps [19:06] <_MMA_> dilomo: And got 7fps. :P [19:06] that's without GPU [19:06] but with GF 8800GTX you get 80fps [19:07] <_MMA_> dilomo: Is your screen > < this big? You keep typing you sentences in half. [19:07] http://en.wikipedia.org/wiki/Convolution#Definition [19:07] sorry ;) skype veteran [19:07] that's a pretty long calculation for the CPU [19:08] i'm surprised a CPU can actually read that!! 8-D [19:08] -.- [19:10] http://blogs.zdnet.com/hardware/?p=3109 [19:13] dilomo, thanks interesting [19:13] np [19:14] the fastest algorithm I can imagine for drawing shadows is a for that prints a pixel each clock and moves to the next pixel uses the alpha value following an incremental logarithmic function [19:15] for blur operations you could follow an algorithm similar to the fake blur beryl plugin: strech an image 0.5x then strech 2x to get the original size with a loss in precision [19:16] yes but it looks like vista's blurr [19:16] no [19:16] neither of them could reach the same quality than using a convoltion filter [19:17] *of using [19:17] not exactly but the blurr of compiz is better [19:17] to my eyes at least [19:18] compiz is using convolution [19:19] but there was for beryl a fake blur plugin [19:19] for older video cards [19:19] aha I see [19:20] http://cgit.freedesktop.org/xorg/app/compiz/tree/plugins/blur.c [19:22] it looks like a complicated process [19:22] <_MMA_> I really don't know why GNOME doesn't officially adopt Compiz. KDE has their own stuff now. Seems like this would be a easy fit for GNOME. [19:23] dilomo, read createlineargaussiankernel [19:23] or similar [19:23] dilomo, it looks like a convolution, through a sum [19:23] _MMA_: probably it would happen in GNOME 3.0 [19:23] remember that you can approximate integrak with a sum [19:24] _MMA_, maybe we will see the clutter backend for metacity [19:24] <_MMA_> dilomo: I don't know. Their still developing their own compositor. [19:24] there's a branch inside ohand's git that uses clutter as a compositing backend [19:25] <_MMA_> Cimi: Hmm.... Interesting. [19:25] http://git.o-hand.com/?r=metacity-clutter [19:25] slow on my nvidia [19:25] but from what they tld me it is quite fast with intel [19:26] <_MMA_> Nice. [19:26] * _MMA_ will bbl. [19:27] Cimi: what is this clutter doing? [19:27] is it like compiz? [19:27] using clutter as a backend means having opengl [19:27] with opengl you can easily manipulate textures [19:27] <_MMA_> Oh. That's a big question. Best to Google. [19:27] <_MMA_> :) [19:27] <_MMA_> AWN uses it. [19:28] and if you get a texture from a window, you can do what you want with it [19:28] just like compiz does [19:28] _MMA_, awn is not using clutter [19:28] netbook-launcher by njpatel uses clutter [19:28] <_MMA_> Cimi: Sure? I thought Neil told ne that. [19:28] tweet (for twitter) uses clutter [19:29] _MMA_, yes I'm sure [19:29] sounds cool [19:29] awn is using cairo [19:29] <_MMA_> Ahh.. Thats right. [19:31] <_MMA_> Ok. No I'm really gone. [19:31] <_MMA_> *Now [19:31] http://folks.o-hand.com/ebassi/tweet-navigation.ogg [19:32] that is using clutter [19:33] looks great. Probably will make more sense with touchscreen [19:34] it's quite fun to use [19:36] that's sure :) [20:07] I have to go bye