(offtopic and quite possibly silly) BeBox architecture, next generation

[Craig Turner <optician%rebel.net.au@localhost>]

Disclaimer: I'll admit now that this is crazy in the hope that you'll
bear with me.

Is anybody here 'into' hardware design? Given the well documented nature
of the BeBox architecture, and the growing momentum of this group, I think
it's a shame that the BeBox is a comparatively slow machine.

What would it take to design a next generation one using more recent PPC
chips? :) The architecture is still pretty up there in its design (as a
complete hardware novice looking at it at any rate :) ) PCI, IDE, SCSI,
and other acronyms associated with modern machines are in there somewhere. 
It's likely that it wouldn't take much to port across drivers for devices
attached to more entrenched PPC projects. The only thing I can see that's
becoming widespread that the BeBox doesn't support is USB. Oh, and you'd
have some difficulty putting together a new IOBoard, but it's possible Be
would come to the table and help out a group of enthusiasts crazy enough
to want to develop a nextgen bebox.  And if they didn't designing a new
IOBoard would probably be simple compared to the rest of the design :)

What would it take? Ask Be really nicely for access to the old designs,
find your chips, determine changes that need to be made to the old design,
based on hardware no longer available, get the boards printed up, stick 
them together, fire them up and then start thinking about a case design :)

I imagine you could get a similar effect much nicer by going directly to
the case design bit and sticking a PC in it, but it somehow lacks the
romance :)

On a related topic,
One thing that I've been thinking about: how flexible is SMP under the
BeBox? Is it a simple switch between processors (not allowing the
architecture to scale up to more processors), or is it dependent on
software and as flexible as other SMP PPC archhitectures?

  - C