Overall I think the idea is a good one. An OS really is not tied to the underlying CPU. Linux is a good example of an OS supporting a number of CPUs. A first step might be to convert all of the SMSQ/E assembly code from M68K to x86. Overall it might be better to use more C and leave the low level stuff for assembly. Granted this might cause some code bloat, but it would allow for greater portability.
I do think the weakness in the QL world has always been the hardware. It's easy to copy software and keep it around for many years. Hardware is eventually going to break. I doubt that a project like this will get started as a group effort. The only way it will come to pass is if one person take it upon themselves to do most of the work and get a working example (Linux was just like this). Once a working example is done, then others might join. Since the source code to SMSQ/E is not publicly available, maybe a good start would be to take QDOS Classic and port it to x86. This would demonstrate the feasibility of the project. Tim Swenson
