Dear KDE developers and designers, I have read your debate about window and windeco themes for Plasma Next. I found many of the arguments used against new styles unconvincing and felt like I had to write a reply. There are at least three issues that underly the opposition to broader changes. One has to do with release cadence, one with the idea of incremental changes and the third with the perception of a lack of quality when it comes to QtCurve. Let's deal with them one by one:
A) The Question of Which Release(s) to Introduce Changes in: The argument by those skeptical of vast visual changes goes as follows: - The first release should show continuity (I believe in the sense of being conservative and trusted). - The second release can have a completely new design and be (in a way) experimental, hinting at the newity of the new Plasma. - By then, the new theme would be tested. BUT: The first release is a kind-of "beta" release, one not for a general audience. This is the very kind of release in which you introduce new stuff that you are not sure of being polished. Why introduce a completely new theme in the second release, if that is the one that is supposed to be generally consumed? Where is the testing happening? In the second release? Or by the few people who choose to change their defaults? I'm not sure this is well-thought out. Even less so, if making a QStyle theme might take "years." B) The Idea of Incremental Changes: The argument about this is: We don't want to shock our users by suddenly changing their UIs to something they might not like as much as Oxygen. Thus, we change one element at a time. BUT: Fist, continuing my point about releases above: The first release is not for general consumption. Screenshots of that release (whether changes are small or large) will be around for many months before a release for a general public is made and by then it's fairly known what it will look like. Secondly, and more importantly, incremental changes are either going to look bad (flat mixed with 3D, gray mixed with the new blue colors, etc.), making all releases until change is achieved look weird in one way or another. And introducing changes gradually does not help test these changes, as they belong together. Crude example: You cannot test a flat button in a 3D Design, but you'd have to make the entire design flat to test the button. And thirdly, artists will have trouble doing incremental releases. They will have to modify the old designs or the new ones (or both) to make them clash less. Changes to the new designs in order to make the transition fell less bad will have to be undone and will be met by criticism from people who got used to these temporary compromises. All this will result in extra work, much lower motivation for artists, weird compromises, a false sense of 'continuity' and a lack of real testing of the new design. C) Quality-related Worries Regarding QtCurve: The main argument here is that QtCurve-based styles may not look as perfect as a QStyle theme and cannot do all things a QStyle theme might be able to do. Thus we need a QStyle theme. BUT: If the claims that a) the designers don't know C++ and that b) making a QStyle-based theme is very hard, possibly taking years) are true, then it will be unlikely that we will see this wonderful group of designers make a new style for all UI elements. Also, it would mean that we won't have a new style until waaay later, meaning not only not for the first or second, by maybe not even for the third or fourth releases. Maybe we won't have a new style at all, but have to stick with Oxygen for much longer. Or did someone volunteer to work on a QStyle for years based on the proposals by our designer? Likely not. This might also mean that we will probably lose some of the spirit the design groups currently has, maybe even the group itself. Why would anyone invest all this time creating, debating, etc. if it will take years to look good? Maybe by then, other styles will be chosen, so we can expect that it is not that unlikely that nothing will come out of the work put in now. The biggest problem here, however, is the assumption that a new style needs to be so complex that QtCurve has to fall short. What we need to do it answer the following questions: - What does a new style need to do in terms of features? What are necessary features for the vision(s) of the designers? - What is QtCurve able to do? - What are the exact shortcomings of QtCurve? What is bad about it beyond personal "feelings?" We need to name things properly. - How big is this ominous "benchmark problem"? Sounds a bit like FUD (though it may not be, but then we need numbers). If the result is that QtCurve can actually do a lot of what is wanted and if the designers are happy with it, why not use the first release that's not meant for a general public to experiment rather than promise slow changes that will look so-so at best and a beautiful design years down the road (if the designers stick around)? After all, the wonderful dynamics of the design group grew out of the ability to play around with things freely. All in all, I think that it's important to use the dynamics of the design group and give it some credit for what it doing. I also think that KDE needs to test the design choices in a meaningful way (i.e. together) and make use of the fact that the first release is not meant for distros to ship to their normal users. Of course, all this depends on whether QtCurve can ever be enough. If not, there probably won't be a new style any time soon. If yes, the designers and developers need to think about how to use it well in order to make a great product. And we might hold it in our hands very soon, rather than years later. Let me add that I believe that I'm sure it's going to be a great release, however you decide on this issue. But the debate was simply too vague and too much based on erroneous arguments, 'perceptions,' and assumptions without any backing up that I thought I try to take it apart. Keep up the great work and thank you for the desktop and applications I have been using for way more than a decade. :) Cheers, mutlu _______________________________________________ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
