Designing Echo icons again

It's been a while since I designed my last icon but for one reason or another I decided I need to make some progress with Echo Perspective Icon Theme again… The final push was me playing with colours—when I had started writing code for Nodoka, I begun with a colour representation. Since we use gradients and shades a lot, I decided to go colour managed and started researching about colour representations and ended up redoing Echo colour palette in sRGB using CIE Luv (in Luma, Chroma, Hue coordinates) as a working space.

The resultant palette is redone from scratch based on the Echo (Perspective) icons that have been already released instead of the current palette for better consistency, but it's "normalized" WRT CIE Luv colour space. The result looks like this:

The right column actually uses normalization WRT CIE Lab colour space, but I decided to stick with the CIE Luv one, which is in the left column. The differences are subtle though.

From there it was ony one step to doing icons, actually. I went all out and decided that I'll

• make more sizes than before to better support current and future DEs, i.e. 16×16, 22×22(24×24), 32×32, 48×48, 64×64, 128×128 and 256×256;
• make similar icons in a single file svg (which means I'll need to write some helper libraries to help split it automatically during build);
• start with action icons in media player, browser and related, which is 23 icons in two groups.

Well, it's a lot of work—23 icons, each in 7(8) sizes, that's 161 pictures. Ugh, that will take a while... Some preview WIP is bellow:

You can notice some design changes, some are new, some have been appearing for a while already, albeit inconsistently. First I went back to slightly darker blue hues than what we were doing recently, it makes better contrast and should work a bit better with darker backgrounds. You could have noticed this design change in my nodoka designs as well (there were three icons in them).

Second change that has been already visible in nodoka designs, is departure from sharp corners. They make the images unnecessary bigger (especially the arrows) and look kinda oldish. I decided that the white outline as well as border should have rounded corners. IMHO it looks better and I hope to retain this style consistently throughout Echo Perspective.

Last change is to shapes—first one I did was view-refresh. The old two big arrows looks sort of heavy and are hard to do at smaller sizes. So I've done something simpler and later I noticed that my new e-book reader uses similar style while loading, lol. My primary source of inspiration in this was an anime I liked for its easygoing style and, believe it or not, typesetting (the anime is called acchi kocchi):

Second shape change was for process-stop icon from circle to the more common octagon. And finally a small change to media-record icon for it to be better distinguished from similar dialog related icons like dialog-no.

Why I'm Still On F12?

First, let me warn you that this is not an objective and exhaustive review of Fedora 13 but just a (hopefully) short list of things (both bad and good) I noticed while trying to use it and to add that most of the issues I'll talk about are bugs/odd features in upstream, so not Fedora's fault.

Evolution

I think it's fair to start with an app that I use most and that is most likely Evolution (especially its e-mail component). The UI is reworked and I'd say it both looks better and is more usable, it's almost pain to return to F12 Evolution's interface. On the flip side, IMAP backend has also been rewritten (or so I've heard) and it hits a dead end here and there and because of that, using Evolution isn't very smooth and joyful. So in short, with future releases there is high probability Evolution will be significantly better that what's in F12, but currently it's one of the pieces that blocks me from advancing to F13. Plus, I forgot to mention, Evolution lost the option to keep gpg passphrase stored during session.

Intel Video Driver

Well, this one's a little bit of pain. I've grown quite used to F12 which has been damn stable so fast, so X or even the whole notebook frozen was a big surprise. Seems the frozenness of whole system have been already fixed, but last time I were in updated F13 (about a week ago), I still managed somehow to endup with half-frozen X — I move with mouse, switch to tty, kill apps (without any graphical response though) or kill X completely. But only full reboot fixed it for good.

Nautilus

Another UI change. Generally it seems cleaner, but what irritates me is that I cannot keep the entry-like input for paths — I seem to be able to access it only via ctrl-l and after entering the path, it reverts back to the button mode. No good, no good.

Webkitgtk

This where the most goodness (from my point of view) went, albeit not exactly as I'd prefer it. Thanks to hard work the guys over at webkit and epiphany, epiphany now can store passwords, but as epiphany does not have sensible session saving (why the hell do I need to kill it to save session?) [and for some other reasons] I use midori, which however it does not seem to know how to save passwords (in web forms, not in web auth pop-ups). The rest of the goodness (I know about) is same for both browsers — html5 youtube flawless support (only fullscreen seems to be missing), html5 video tag finally has handler, sites (like anidb) which serve pages in compressed format are now supported and java applet (I've tried only the openjdk one) is almost working.

Gnome-keyring

In one word it behaves strange. Sometimes the keyring that should get unlocked at startups refuses to unlock, sometimes I get asked for new password for 'default' keyring, which however never gets created… This is incidentally also the last thing that shows when trying to start up gajim. The good thing is that thanks to this I started using emapthy with a chat theme from one of my favourite anime durarara >_<

Conclusion

Well, there are still some hick-ups that keep me from transferring from the probably best fedora release so far (F12) to F13 to feel the new awesomeness, and one thing to keep in my mind is: don't be lazy and finally fill the bugs you've encountered!

Math, anime and my (crazy) head

Being an undergraduate (I've recently completed my Bachelor's degree and am currently starting with Magister degree studies) in physics I tend to take interest in things where math is concerned, even when it is just a snapshot of a math test I notice during watching anime... A few moments ago I came across this:

The title roughly (my Japanese does not go much further beyond being able to look up kanji and some basic everyday-life phrases) translates to "First year, first semester, midterm exams, math I.", the first problem is something like "Collect similar terms and put them in order in the following expressions." and the fourth problem goes like "Factorize the following expressions." Well, from this I can guess that the math level on first year of 高等学校 (high school, students are around 15 years old) in Japan is roughly the same as in the first year of střední škola (high school, students are as well around 15 years old) in Czech, although I've never seen such answers to these types of math problems... Well, it's from anime and it's quite usual there to go a bit overboard, but nevertheless it gave me a good laugh.

On a similar note, over the past week from time to time I've been trying to find various derivations of Lorentz boosts in general direction in hope to come up with an elegant one. Today I've finally managed to find one that is purely mathematical and based on the assumptions that a) general Lorentz boosts does not contain space coordinate rotation and b) Lorentz boosts in a given direction form a one-dimensional sub-group of the Lorentz group (which is isomorphic to O(1,3)), whose another sub-group is the group of rotations O(3). Lorentz group is a group of (linear) transformations that preserve the Minkowski metric

Limiting only to transformations that do not contain time or space inversions (and thus in matrix representation they have determinant equal to 1 and the "time-time" component greater than 0) I get a continuous subgroup SO(1,3) and can further derive limitations imposed on the matrices forming it's Lie algebra (so(1,3)). It turned out, as expected, that the 4x4 matric has a 3x3 antisymmetric submatrix (the space rotations), has zero trace (because determinant of the group representing matrices is 1) and the remaining elements are symmetric – which also leads to (already known) fact that the dimension of this group is 6 (three space rotations and Lorentz boosts in three mutually orthogonal directions). Note: I use geometrized units in which c=1 and otherwise the remaining elements wouldn't be symmetric, but differed by a factor of c^2.

From there, finding the matrix describing the Lorentz boost in general direction is simply a matter of filling the symmetric elements of the matrix representing the so(1,3) algebra (and leave the other elements 0) and making an exponential out of it. Further one can replace the coefficients describing the direction of the boost and hyperbolic sines and cosines by the usual gamma factor and three-velocity components to get the usual from of the transformation.

I really like this derivation – even though it's rather long compared to other derivations I've found or came across during past week, it's purely mathematical in nature and does not require knowledge of the Lorentz transformation for frames in standard configuration.