April 14, 2003 by Allen George
Graphics Under Linux
Usually I would spend the remaining 1/2 an hour till bed time trying to make the most out of my time. However, today I have decided to try a different approach. I often talk since it helps me organize my thought processes and quantify the ‘swirl of ideas’ that I experience. Now, instead of monopolizing someone’s time, I will use a passive medium that should serve the same purpose.
Two days ago, xwin.org was created. I had understood it to be a fork, but after careful reading, I believe I am wrong. Xwin.org currently exists as a ‘community’ site. My personal opinion is perhaps more cynical – I believe that xwin.org simply exists as a place to ‘blow off steam’ by the community. A first sampling of articles by anon. contributes includes things that everyone interested even tangentially in Linux graphics talks about. Finally, the questions that people have been ruminating over come to the fore.
Unhappily however, the site is not about the following:
– It has no code
– No move towards moving towards a new version of the X protocol
– No move towards completely deprecating or even dropping unused (little used) extensions from the implementation.
– No move towards creating a separate driver subsystem independent of X.
The core X developers are of the impression that the X protocol, while old, is more than adequate for today’s and future needs. Although this is understandable due to the vast body of applications that conform to the X11 standard, I believe it is evidence of short term thinking that they are unwilling to acknowledge that perhaps, we need a way forward beyond X11. The greatest truism in computers is ‘beware backward compaibilty’. It is good business practice, and improves matters for end users, but pevents true improvement and achievement.
The X developers have neglected one thing in their support for X – the reactions of all the endusers. People want something better. It is as simple as that. Most, like me, having converted, are impressed at the stability, the throughput and the speed of a Linux installation, but find ourselves underwhelmed when using any graphical application. An honest review of normal 2D applications between Windows and X reveals that X _feels_ slower. When people make these statements they are immediately asked to “demonstrate proof”. How does one demonstrate a feeling? What the X developers are ignoring however, is that by comparison, no one claims “replace the kernel”, “replace Apache” or “replace GCC”. Why? Simply because these projects work and they work well. When end users see tangible benefits and improvements, fast releases and a community that is not afraid to improve and extend, they are silent.
There’s also the future to consider. OSX and Windows Longhorn ae the future of desktop GUIs. Compared to them X _is_ dated. My question to the X developers is “Can you agree to this?”. These GUIs will bring 3D (accelerated via the GPU) to the desktop. OSX at this moment takes advantage of the capabilities of the GFX card for certain operations. X does not seem to be rising to this challenge.
My time is up today – tomorrow, I shall continue and write more of what I know.