8 Jan
1999
8 Jan
'99
7:11 p.m.
|Yes, it should have. And the more MS waited to
improve DOS (because there
|was no incentive since they already had a monopoly, or because they were
|sloppy programmers, or for backward-compatibility reasons) the more
|excruciating the thought of improvements which would break old programs
DEC pulled it off with VMS. The people who had to
have the old compatibility just kept running the
old hardware and software, and everyone else moved
forwards.
And right happy about it most of us were. We ended
up with an awesome development system that did OK for
realtime (VAX/VMS) and an awesome realtime system
that still needed better developer tools (PDP/whatever)
That may have worked, but the factors didn't come together the right way.
There were many well-designed 8-bit machines and many manufacturers who
could have conceived of a plan like that. Unfortunately the innovative
manufacturers (Apple, Atari, Acorn, maybe Commodore) generally profited by
working as close to the hardware as possible. Using your model, their
analogues of the PDP-11 -> VAX evolution were a lot less clean and neat.
Note that there are some assumptions about the 8-bit period. The CPU's were
not as fast as the PDP-11 was, and people apparently didn't want to pay the
high price for memory management. Also, there's the matter of games (which
make great demands on small CPU's and don't lend themselves to backward
compatibility). How does that kind of software fit into your hypothetical
situation? What about on the real PDP-11 or VAX?
The only company that _did_ make a serious effort at backward compatibility
was DRI. Of course that's bitterly ironic -- they may have won that battle
but they lost the war. In our universe, IBM ruined DRI's approach by
designing crappy hardware and writing a crappy BIOS for it; then Microsoft
entered the picture with its crappy DOS that incorporated a different
perspective about the hardware than IBM's; then programmers started
violating all the rules to get decent performance. And to rub it in,
neither Microsoft nor IBM seem to have the innovation that the previous
generation of manufacturers had.
I think the only way your scenario could have worked would be for some
hppothetical company to have had a well-designed (made for future growth)
16-bit machine early in the game. A further irony is that DEC was such a
company, but didn't sell the fruits of their labor. I think some other
company might have worked, though. Intel? Corvus? Xerox??? Really, I
don't know if any company (in our universe) fits the bill.
VMS was at least in the works, and UNIX was around in
some forms. There was precedent. But the guy hacking
DOS was just messing around, trying to do a slightly
better CP/M (as was Digital Research, who GAVE AWAY the
whole stinking computer world to Bill Gates on a silver
platter, with a military escort, and a ticker tape parade.
A lot of people want to read Gary Kildall's memoirs... if the true story is
written down, it's gotta be in there. The problem is that they have never
been published.
Not that I'm bitter, you understand. 8^)
Well, you COULD fight the trend and live in the past (like a lot of us do)
by collecting classic machines. You could also start designing the next big
machine and have it ready for the time when Microsoft weakens.
-- Derek