March 18, 2007

Software is more reliable than people!

Posted by Ben Simo

Software is 100% reliable. It does not break. It does not wear out.

We can depend on software to exactly what it is coded to do every time it does it.

This is why software quality stinks!

The consistent repeatability in software is both a blessing and a curse. We can depend on software processing the same data in the same way every time. We can also rely on software to not do what it does not do every time.

The repeatability of software is both its greatest strength and weakness. A simple mistake in design or implementation will forever repeat itself each time a computer program runs.

Over 50 years ago, three mathematicians wrote:

Those who regularly code for fast electronic computers will have learned from bitter experience that a large fraction of the time spent in preparing calculations for the machine is taken up in removing the blunders that have been made in drawing up the programme. With the aide of common sense and checking subroutines the majority of mistakes are quickly found and rectified. Some errors, however, are sufficiently obscure to escape detection for a surprisingly long time.

[R.A. Brooker, S. Gill, D.J. Wheeler, "The Adventures of a Blunder", Mathematical Tables and Other Aids to Computation, 1952]

The source of the problem is people. We are not reliable. We make mistakes. Software amplifies and repeats our successes and our mistakes equally well.

When our software encounters the unexpected, errors occur. As developers and testers, we need to expect the unexpected. Think about how requirements might be misunderstood and clarify any ambiguity. Think about how users might misuse a system -- either accidentally or intentionally -- and ensure that the system can handle that user behavior.

The software we use today is exponentially more complex than the software being developed 50 years ago. There have got to be more opportunities for blunders in today's software than there were in software development half a century ago. And with that complexity some errors become even more obscure and are more likely to escape detection by developers and testers. Users, however, seem to easily encounter these errors.

A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.

-- Douglas Adams