Justin, Good thoughts. We gave this some time as well. The biggest problem I see (which is one of Linux's strong points) is the dynamic nature of the kernel. If we spend a great deal of time now pruning, we will likely have to reproduce the work later to move to an upgraded kernel.
Personally, I don't think it matters what is "under the hood" of the OS. I want to try and find the easiest to manage and most flexible. Granted, contradictory goals usually, but I think Plan9 offers the best hope at this point. But, I can be convinced otherwise. I was sure wrong about eCos! :-) Bill At 03:28 PM 5/20/2002, Justin Cormack wrote: > > > > rather than build a mini-os, I wonder if we shouldn't start with the 1.0 > > linux kernel? > >I really dont think 2.4/2.5 is beyond all hope. With nothing compiled in >it is 218053 bytes. But I think too much is being linked in despite not >being selected. A few weeks at least seeing what can be removed would >potentially give us what we need. > >The main problem I forsee is that there are various things that people wont >make modular for various (good) reasons, but could be compiled out or >replaced with minimal versions. > >I dont really want to use plan9/hurd as my bootloader unless I am going to use >it as my OS or the multiple support issue comes up again. And they are >minority platforms. > >100k (or even 50k) off Linux should give enough for 1 or 2 boot methods >available in linuxbios. > >Justin
