Re: Schneier, Data and Goliath: no hope for privacy

In article <vpo4uc$2omvt$1@dont-email.me>, Rich  <rich@example.invalid> wrote:

I had (if memory serves) at least one Pascal class, one Fortran class,
and an assembler (CDC Cyber 7000 - a really weird CPU on the inside)
class, all required classes for Engineering.  Pascal class was trivial
(had already done plenty of UCSD Pascal on Apple II in high-school) so
just had to adjust to the small difference in the CDC Cyber Pascal we
were using.  Fortran was similarly trivial, but oh did I come to hate
Fortran in the end.  Just had to learn the "fortranisms", as I already
understood the over-arching "how to program" aspects.  The assembler
class was also itself trivial (had done loads of 6502 assembler by this
point, and some 8086 assembler, provided one considered DOS's debug an
'assembler' of sorts).  Just had to "learn the language" rather than
the "how to program" part.

That's pretty unusual.  The reason why Fortran is a good thing is because
engineers can't be trusted with pointers.  And COMPASS?  That's a very
very strange assembler to teach.... I went to gatech which had Cyber
machines which the CS folks avoided like the plague.  COMPASS is not
exactly a normal assembler and has a lot of fast-float-performance
craziness... it is not something I'd really teach anyone whom I was trying
to teach about the principles of computing or how systems work.  And the
PPUs code?  That's worse than IBM channel controller stuff.  I'm sorry
you had to do that.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."