I don't understand your last comment. The source code is public, development plans and status are discussed constantly on the mailing lists & IRC channels, and anyone looking for help getting into driver development gets enthusiastic support from the other developers.
The Linux graphics stack is going through huge changes right now, and those changes constantly introduce new driver issues, particularly for older drivers that don't get a lot of attention otherwise. Most of the problems have been tracked down and fixed, but others are still showing up. Some of the crashes take weeks or months to solve, and there are limits to the amount of time developers can spend.
There may be a lot of older systems around but I think you'll find that usage drops off dramatically after ~5 years or so. Developer interest also tends to fade once the hardware gets old enough that it can't support the new features they're working on (and the r2xx family is at that point now).
Bottom line here is that if you want to keep using an older system (which is fine) you really have two choices :
1. Keep updating to the latest distro & driver versions, accepting that there *will* be regressions from time to time, that you will need to do a fair amount of investigation and experimentation (including filing good bug reports and working with developers awhen they have time) to have a chance of getting them fixed.
2. Pick a distro version that works well and stick with it. If you go with something like a Ubuntu LTS release you can continue to get critical updates for a long time but without the constant changes that cause you grief.
If that was my laptop and I wanted trouble-free operation I would install Ubuntu 9.04 and stick with it for the life of the system. There will continue to be new features added for a while (including some neat things like KMS), but you also need to understand and accept that those new features will bring new problems as well, and that for older hardware there are not a lot of users and developers to find, investigate and solve the problems before you encounter them.