Re: Don't discredit OVC with falsehoods!

From: Karl Auerbach <karl_at_cavebear_dot_com>
Date: Fri May 07 2004 - 12:23:18 CDT

> Not that I want to argue this further but...By the way she did get the
> flavor of the linux "=" incident correct I believe.

Good design rules help with things like this.

For instance for C code, use of "=" within a conditional could be
prohibited by a design rule - and tools could check for the use of such
constructs.

Similarly, it is important to use the facilities that modern languages
provide, such as the "const" keyword in C and C++.

Liberal use of assert() tests in code is also a good thing.

Things like those above are easy and inexpensive.

Some things can be hard and expensive:

When we did formally verified software we had to make formal assertions
about side effects of every function. (And we didn't even try to make
assertions about behavorial signalling kinds of side effects, e.g.
modulation of paging as a signal to other processes.) Then we had to go
through a lot of work to actually prove that the code met those
assertions. Needless to say this was slow, usually boring, and very
expensive work.

                --karl--
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Mon May 31 23:17:22 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT