Mailing List Archive

Mailing List: techdiver

Banner Advert

Message Display

From: <devon@ol*.ch*>
Date: Thu, 3 Aug 95 16:17:43 +0200
To: cavers@co*.ci*.uf*.ed*, techdiver@terra.net
Subject: Re: Rebreather experience
> You are not completely correct in this statement.  Whether or not software
> can be certified as bug free depends upon what context you define bug free.
> It is quite possible to *PROVE* that software is correct

It's possible to prove formal algorithms but not software. The development
of software relies on tools. Language specifications have ambiguities,
compilers have bugs, debuggers have bugs, etc, etc. And you can't prove
those because you'd need other tools to do it! Proving algorithms is more
for the theoretical world than reality.

Even the US military (who have a large interest in correct software but
not as large as the rest of the world!) runs different redundant versions
of their software for reliability. For example, they might run the same
code on two different hardware platforms compiled with two different sets
of tools to spread and "thin out" the errors. Or even two copies of the
same program written by different groups.

Trusting algorithms and trusting software are two very different things.

Devon

Navigate by Author: [Previous] [Next] [Author Search Index]
Navigate by Subject: [Previous] [Next] [Subject Search Index]

[Send Reply] [Send Message with New Topic]

[Search Selection] [Mailing List Home] [Home]