Do you think
that tools will not evolve in 20 years?
Short answer: No. ;-)
Heck, you can't get schematics, technical information, or anything on most
newer hardware now, so one would need to reverse-engineer almost
everything; and the tools that do evolve that are necessary to work on even
today's stuff (fast oscilloscopes, etc.) are priced out of the hobbyist's
pocketbook.
I think it's worth considering if the necessary tools to
non-destructively reverse-engineer a modenr machine exist now _at any
price_. I think they do not.
If you think back 20 years, logic analysers existed. They were beyond the
reach of most, if not all, hobbyists, but they existed. Now, of course,
many hobbyists have one. Ditto for a 'scope. I think it's very likely
that the hobbyists of the future will have fast 'scopes, BGA rework
facilities, and the like.
But there's more to it than that. Suppose you have a modern PC
motherboard with some ASIC in the middle. You will be able to remove and
replace it. You will be able to connect it into a test circuit. But I am
going to assume that such an ASIC has many bits of hidden 'state'. I also
don't believe there will be any more laws of digitial logic discorvered
(I think that subject is pretty well understood now), so there'll be no
new way of determining said 'state' from the external connections, or of
forcing the chip into every possilbe state, or...
So, what aew you going to do? Risk uncapping the package and looking at
the die? Will you be able to make anything out of it? (speaking from
experience I know that a large board of discrete transistors ia about the
worst possible thing to have to reverse-engineer, you have no idea what
the various sections do, whereas, say, a microprocessor pretty much ties
things down. Soing an ASIC would be like the former, only worse). Maybe
there'll be CAD tools that will help, I am certainly not holding my
breath, since I've yet to see _anything_ available today that can really
help.
And uncapping the package will certainly risk damaging the IC. Would you
want to risk doing that on what is a one-of-a-kind machine? Think of
doing the HP01 watch _now_. It's a single ceramic hybrid module inside
contiaining 5 or so silicon dice. The watch is rare. It's also reasonably
well understood, since it's knwon to a be a 48 bit version of the 56 bit
chipset used in HP calculators of the time. But if you had one would you
want to ceack open that module to reverse-engineer it? I wouldn't.
-tony
