Thread: Future of Litestep
Tue Nov 23 2004, 01:41PM
Guest
Hello,
The end of the CPU-driven user interfaces era draws near,
since most computers have this fancy graphics card sitting there pretty much idle most of the time.MacOSX's gui has been rendered through OpenGL for years now, and Windows is about to switch to Avalon - which draws the user interface using DirectX.This is a logical step, with exciting possibilities. Question though - where does that leave our beloved shell, Litestep? Since there's no uniform 'draw'-API as i understand it, it's left to each module to draw it's graphics itself, using outdated GDI tech.
Will Litestep evolve with this? Uniformize draw functions and pass them on to the Avalon API? I feel (and hope plenty of LS users agree) that LS will benefit immensely from this, from a resource point of view anyway. What do you people think?
Thread: Future of Litestep
Tue Nov 23 2004, 02:11PM
Registered Member #26
Joined: Fri May 14 2004, 05:32AM
posts 199
If M$ sufficiently opens up the new API, alternative shells will use it. The coolness factor will draw them in, along with lower CPU use. Listed just to keep it neat.
1. "For years now" specifically refers to OSX, which isn't much older than Windows XP. Also, the new GUI will be using DirectX. Guess what? So does the old one. #2 and #3 are the issue, not pure CPU vs. pure GPU. So far, I haven't seen much interesting come of it...are there any USEFUL things the new GUI does?
2. Microsoft is afraid of OpenGL, dispite needing to have options for it. DirectX is a good way to go for gaming, but for general use, it only cripples. Every decision since the first XP release makes them look very much like they just don't get it, in general.
3. Most alternative shells, including LS, are now open source to some degree, and plugins/modules tend to be , as well.
4. I believe that a common display module will be necessary for good future growth. A common keyboard/mouse API wouldn't hurt, either.
5. All of this, however, is not as big of a hurdle to progress as not being able to define objects is. Instancing in xLabel and xTaskbar is great, FI, but requires tons of work for what is mostly duplication. Unfortunately, with run-time hooking out of the picture (that would allow defining and then creation via scripts), there's no good way to deal with this, and it adds weeks, if not longer, to theme development time, just testing things out. Exactly how this would be dealt with at this point, I can't say, except that scrollbars are OK until I try to make several different kinds of them, each marginally different from the others [ Edited Tue Nov 23 2004, 02:14PM ] [ Edited Tue Nov 23 2004, 02:16PM ]