>> For example, let we observe Xorg consolidation, based on X Window System. >> This source code tree evolved over the time in such different ways that >> now it comes to the critical point where it couldn't be fixed anymore. > > I have no idea why you think X can't be fixed anymore - we're making > improvements to X all the time - X has improved far more in the last > few years than it did in the 10 years previous.
Certainly you did. But I think that X Windows System has some bad solutions regarding software design on a high level, that you/we could improve it in next 10 years but we will never acheive level of design as Sun's engineers did in SunOS. X Window System has bad architecture - that is all. > >> What we need? We need some Senior Developers conference to start to think >> how to make X Window System more comfortable for programmers, > > X.Org made a huge investment of time here in past years working on this > - converting from the little used imake to the widely used GNU autoconf > & automake, breaking the build from one giant tree into hundreds of > individual modules, so that a developer who wants to build a new ATI > driver module doesn't have to build all the libraries, clients & fonts > in the process. This has been part of what's brought about the > above mentioned resurgence of X. > Is it possible instead of GCC to use Sun Studio 12 toolchain? >> even if that means developing completely new Windowing System. > > Many attempts to develop a completely new Window System have been > proposed. None have attracted enough developers or applications > to be more than a curiosity. A completely new window system is > a project of huge proportions, and if you're not compatible with > X, no one can run their software or use their hardware. > >> We need more organized architecture, well integrated yet modular, >> to allow that each hardware vendor write just one object module to serve >> as a driver for each Implementation of X Window System > > I thought that's what we had - each hardware vendor can provide a > single .so module for their Xorg driver, such as intel_drv.so or > nvidia_drv.so - what would you change from that? Or are you > unhappy that they have to provide both Xorg & kernel driver modules > if they want to support functionality like suspend-and-resume or > 3D hardware acceleration that needs kernel support? Bad 3D graphics card support is just one of consequences of bad architecture. It emerged since GPUs become increasingly powerful in recent years. Gallium3D is one ad-hoc answer to this problem. We all hope it'd be successful. What will happen when we have heterogeneous microchips with many different cores in itself (scalar, superscalar, etc.)? We should build again new technology to deal with it. This is my point of view. I hope I was clear. Uros > > -- > -Alan Coopersmith- [email protected] > Sun Microsystems, Inc. - X Window System Engineering _________________________________________________________________ Show them the way! Add maps and directions to your party invites. http://www.microsoft.com/windows/windowslive/products/events.aspx _______________________________________________ indiana-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
