but this didn't mesh with my own experience in kde. i didn't get the impression of 1500 different actions, applications and mimetypes. certainly i could picture 1500 items in the ui, but are they all unique? hmmmmm...
so i wrote a quick script that scoured my hard disk for all the icons used in kde and the various kde applications out there that i have installed (a rather large number) and put them all in one place without overwriting icons with the same names. i then looked at them in konqueror and it quickly became obvious that we had a huge amount of duplication:
- visual homonyms: icons that looked the same, even if they had different intended meanings (actions) behind them
- semantic homonyms: icons that have the same or similar meaning, even though they may look different
- plain old duplication
that last group is an interesting one: it's the result of application developers copying icons from other apps or even kdelibs itself and renaming them. why? various reasons: because the icon isn't in kdelibs, the icon is named after a different action, the application developer was using it as a placeholder until a "real" icon for their app was made (which often doesn't happen). the correct course of action is to either move the icon into kdelibs or not rename it if it is. in any case, with kde4 approaching we can consolidate a lot of this duplication. the homonyms are also cause for a huge amount of duplication.
so how to get a handle on it? i fired up digikam and started tagging. i did most of the tagging on slow evenings in paris and on the airplane flights.
doing this tagging is quite enlightening. over 70 tags so far (though more could be employed, and by the end of it likely will) and it requires at times a knowledge of the names used in the code for actions, usability principles and even a bit of icon art appreciation ... the results will get passed on to the artists who, i think, will be relieved to know that they have far, far fewer than 1500 icons to make.
as application developers, we should really try not to recreate this mess over the next 5 years as we did over the past 5 years.
10 comments:
I wonder how you manage to always pick up those points which I think are important to do. (I also wonder how you manage to do everything and be good at it.)
For me as an icon set maintainer, this is actually great news. Side question: Will the information that you find made public? Some sort of a "Which icons do I have to make for a complete KDE icon set" document would be awesome.
Btw, any news if KDE will use the tango icon naming specification or rather stay with the current broken naming scheme?
is it possible to see the latest oxygen icons somewhere?
as was already mentionned by jakob petsovits, isn't the tango icon naming scheme already solving this problem?
@jakob:
> I wonder how you manage to always
> pick up those points which I
> think are important to do.
psychic link, obviously ;)
> Will the information that you
> find made public?
yes ... i've completed the tagging, am now waiting for some feedback from artists ... if you'd like to participate in this part of it, please email me (aseigo at kde dot org) and i'll send you a url to the finished results thus far.
> if KDE will use the tango icon
> naming specification
you mean the freedesktop.org icon spec with the naming addendum? yes.
(calling it the "tango icon naming spec" is like calling the desktop file spec "the kde ini file spec")
> is it possible to see the latest
> oxygen icons somewhere?
they are in kde's svn right now.
> isn't the tango icon naming
> scheme already solving this
> problem?
no. the problem is how applications use icons in kde. changing the names of various icons wouldn't actually fix anything here; the icons covered in the naming spec aren't the ones that have the duplication issues. so even if we'd had a common naming spec 5 years ago, we'd still have this current problem.
it is due to applications whose developers figured had "special needs" and so created variations on or straight-out copied icons over and over with little cooperation between the developers and artists resulting in a huge number of icons with lots of duplication between them.
> > if KDE will use the tango icon
> > naming specification
>
> you mean the freedesktop.org icon > spec with the naming addendum?
> yes.
>
> (calling it the "tango icon
> naming spec" is like calling the
> desktop file spec "the kde ini
> file spec")
Well, of course you're right, but in this case the wrong term accounts for, don't know, rough estimate, about three times more readers that wouldn't have got the context otherwise.
Or maybe I should just quit thinking that people are as uninformed as I expect them to ;)
Sorry for such an off-topic comment, but since you're in charge of Plasma: have you seen what Novell has done with the Gnome menu? I think something similar would be great for KDE 4. Just a suggestion.
> have you seen what Novell has
> done with the Gnome menu? I think
> something similar would be great
> for KDE 4.
Something like that is already in work and researched by Celeste Paul ("seele") and will definitely go into KDE4, as far as I'm informed.
Am I the only one who can't stand the Crystal icon set?
Probably...
> Am I the only one who can't stand
> the Crystal icon set?
no. though most people don't mind or even outright like it. ;)
it is showing it's age though. worse thing is that the original artist took off and decided he'd like to control crystal rather than support the open source project he created. he does have a nice design studio because of this decision (or in spite of it?), but it resulted in N people over the last several years just adding random stuff to it that doesn't really mesh.
for kde4 it's all moot anyways. oxygen is one its way.
Post a Comment