Notes
367_22
To Prof. McCarthy
1. From what medium come the developers of operating systems management
programs (mathematicians or engineers)?
2. What courses in operating systems are given at universities?
3. What is the productivity of a system programmer?
4. Your opinion on IBM/360 system and its future?
5. Your attitude to OS/360?
_____________________________________________________________
367_36
Prof. McCarthy!
What is the way of reassuring that the adopted computation method
doesn't lead to a great precision loss?
_____________________________________________________________
367_38
To
Prof. MacCarthy
Wouldn’t
a programmer wishing to formally
verify his program become similar to a
centipede: as is well known,
the latter, “algorithmizing” its
activities, became unable
to move.
At
first, yes, but we will learn how to
formalize
________________________________________________________________
367_50
For real progress) such things as syntax-directed <...> and Uncl systems
will be of great aid. And I agree that the fact that previous attempts
which failed in Uncl systems should not be discouraging since they were
attempted at a time when much less was known.
However for real progress we need proofs of the correctness of programs
as suggested by McCarthy. The problems involved seem to me difficult and I do
not believe that real progress will be made in 70's.
We learn what we want as we program.
Back
|
