RISC was the way to go almost 20 years ago with the likes of the BBC`s Archimedes computers.
Their problem was the same problem every1 has ever face when producing an alternative to x86/m$ platforms.
At the time the Archi computers were very very good, with a good number of practicle applications used by schools in the UK (as part of a deal) and some businesses, but only a few "quality" entertainment packages, usually from the mainstayers of the bbc micro legacy. And of course u can play doom and quake on them.
The BBC ditched them (making computers) after a while and Apple started to move from the 68k to ppc motorola classes - which similarly are far more powerful than the equiv x86/m$ platforms.
By the time Apple went G3/4, u could forget any thoughts of competition, the x86 was/is being clocked faster and faster, with not really any great advances prior to things like mmx 3dnow and now cores. Just get more cycles with more transistors - as hot as u can, don`t worry about memory.
The motorola group of processors, as any1 who`s used them seriously can tell u, are and would have been a more beneficial way for the computing world to go. They don`t need cores to multitask for a start. They`ve been capable of multitasking since Amigas used them, we`re talking decades ago. Now the adverts on u`r tvs are about multitasking with u`r pc.. do several things at the same time... well duh! we could before.. but we didn`t and don`t need 20000gigs of fast fast fast memory to do it.
Risc is not a part of my assembly vocab, I pretty much got to the advanced 68k processors. Probably missed out on alot.
I want an Apple. >.<