% fortune -ae paul murphy

Reformulating Murphy's law

As everybody knows, Murphy's law says that what can go wrong will go wrong - and lots of people will be familiar with the public policy version of this known as the law of unintended consequences: that policy implementation often produces results opposite to those expected.

As normally stated Murphy's law isn't really a law as much as it is an observation in that it describes what is seen to happen, rather than stating a principle whose application predicts what happens.

State the law as a law, however, and while you lose the zing you get from the traditional cynical observation about reality, you do get a much broader picture:

If real actions are taken on the basis of beliefs about reality, then their consequences will align with those beliefs to precisely the extent that the beliefs align with reality.

Notice that this predicts Murphy's Observation: go into something with unrealistic beliefs and you're setting yourself up for failure.

The law's operation is most obvious with respect to the contrary consequences so commonly seen in public policy: where peace policies lead to war, climate hysteria produces economic failure, short term dietary fads produce long term disease, and nationalized health care raises costs, limits access to care, and reduces life expectancies.

The applications to IT are less obvious but equally compelling. If you believe, for example, in the achievability of wintel security, you'll spend lots of money setting your employer up for an embarrassing loss of customer data; if you believe Wintel cheaper than any Unix you'll spend lots of money causing your employer to lose the cost and productivity advantages of the cheaper technology; and if you buy into the idea that Wintel empowers employees you'll find yourself spending lots of your employers money restricting what those employees can do with "their" machines.

Or, more traditionally, if you go into some software project or demonstration hoping it will work; it won't - but if you align your beliefs with reality through clear thinking, good programming, and rigorous testing, it will.

As one example of Murphy's law at work, I want to consider one of the things going on with Sun's market. Both HP and IBM are doing the Sun is dead wink nudge dance with customers and urging them to buy their own, much more expensive, products instead. That's good for them - but really bad for your employer if you're eager enough to believe this to fall for the pitch.

The problem is this: they might be right, the financial market attacks on Sun were crippling and customer reaction is hurting as much again, but the bottom line is that the facts aren't in yet and it's simply premature to draw conclusions one way or the other.

Thus the guy who sent me an email last week gloating about Sun's death and proudly attaching a copy of IBM's press release claiming to have set a new SAP processing record with a p550 was acting out a fantasy - and if he succeeds in getting his employer to shell out for a couple of the things the bottom line on Murphy's law is that the results are going to align with reality, not his preferred beliefs.


The IBM press releaseis carefully couched to limit comparisons to other eight core systems (i.e. to x86) - because the Sun 5440 with Oracle beats IBM"s system by about a quarter on performance, at something significantly less than a third of the dollars.