John Foust wrote:
I think this is crucial for those who wish to
understand why the Amiga
didn't live forever: although it was 32-bit, its architecture had a deep
reliance on wide-open shared memory, preventing it from reaching the
next level of robust behavior. Tricks like "Enforcer" certainly helped
developers write safer code once the 68020 came along, but it was too
little, too late.
I knew there was a reliance on shared memory, but I didn't know it was that
deep. Are you talking just about the custom hardware? (When the Amiga was
new, Amiga clearly had to work its butt off it get the chips to perform so
well -- making them squeeze through an MMU while doing DMA was obviously not
practical. Do you know if they even had plans to add memory management?)
Or does the OS rely on shared memory as well? (by being nosy about its own
routines, perhaps)
I'm tempted to say that after its initial success, Amiga rested on its lau-
rels and didn't try to improve the hardware for a long time (and they only
sort-of tried to improve the software). I don't know if it's *true*, but
it's *tempting*. Then they got around to the AGA graphics hardware which
is much better, except they were never able to catch all the bugs!
Now they want to combine the old architecture with fancy new CPU's and hope-
fully memory management -- good luck but it sounds like a lot of work.
Do you know if they're trying for software compatibility?
There is another lesson here, I think. It's one reason why adaptation has
been or will be so hard -- ANY decision you make about a system, good or bad,
especially bad, will affect the system for longer than you think.
-- Derek