9 Dec
2010
9 Dec
'10
2 p.m.
On Wed, 8 Dec 2010, Tony Duell wrote:
Sure, but the patents help identify signal names and understand the
priciples of the machine. You can't easilly find out how a machine works
by just looking at the flip-flop or some random logic. We had to repair
Well, maybe some people can't, but there are plenty who can :-). More
seriously, this is really what reverse-engineering is about, not just
tracing out the schematic, but understanding what it is saying.
No, I
haven't. The objective is to make a precise gate-level simulation
of the calculator, both for understanding and which can be used to
track down faults in a broken instance of the real thing. Patents
generally don't provide the level of accuracy for that objective, so it
generally doesn't occur to me to look to patents for these purposes.
As n aside, some of the HP desktop calcualtor patents are very detailed,
and include sechematics, commented firmware source, extension ROM
sources, and so on. But as you imply, there are often suble differences
between the machien described i nthe patent and the actual prodcution
model. I regard these patents as a very useful resource and well worth
reading, but you need to check against an actual machine. some Friden 132 boards just a few weeks ago, the block
diagrams and
descriptions in the patents were quite helpful.
I think we're saying the same things. Certainly for the HP machines
(which I am more familiar with), the patents are very useful. They will
help you understnad what is going on in the real machine. However, if you
attmept to repari a prodcution machine using the patent as a 'service
manual' with no more thought then you will run into problems. The machien
in the patent and the actuall machine are not quite the same. If you use
the patent intellegently, examine the actuall hardware and see how it
relates to the descriptions and diagrams in the patent then you'll sort
it out.
-tony