28 Aug
2006
28 Aug
'06
9:38 a.m.
On 8/28/2006 at 12:01 PM Dave McGuire wrote:
Either that or let's change it to the
"twenty year rule" and revisit
it in another few years.
When we're talking about old software, there's not a heckuva lot of
difference between a Pentium or a 486 (and clones) or even a 386
(especially when the math coprocessor is installed) or a 386SX. Windows 9x
runs on them all just fine. Other than the PCI bus not being present on
the 386 boxes, the distinction to me would seem to be somewhat arbitrary.
And the 80386 is almost exactly 20 years old. As far as software is
concerned, the OP could well have posted a question about an 80386 system
and gotten the same answers.
Okay, so let's turn back the clock 5 more years and set the cutoff at 25
years to 1981. That would let out many of your favorite workstations, IBM
PC-XTs, all VAXen but for the 11/1780 and the 11/750, the Next boxes and
even a few CP/M 8-bitters. Conversation on the list would slow to a
trickle, methinks.
It's interesting that the computer field seems to model a biological
system. There's a tremendous amount of species diversity in a young forest
of, say, Douglas fir. As the stand matures, the heavy shade cover forces
out most other plants until one is pretty much left with Douglas fir and
various mosses and ferns that can tolerate the shade.
Right now, we're looking at what appears to me to be a mature technological
system--we've got nothing but a forest of Pentiums and Pentium look-alikes
with other minor CPUs occupying the role of embedded support functions.
You've seen one, you've seen them all. Short of a major upheaval, I'd
expect things to stay like this for a very long time. And it's boring.
We could exclude certain CPUs, but that doesn't address the issue that
anything that runs on an older CPU will probably run on a newer CPU of the
same family, so one might as well include the newer ones.
I don't know what the answer is.
Cheers,
Chuck