Code thieves and the people who hire them

I saw in the news that there’ve been two cases lately of programmers stealing source code on the way out the door from their employers, presumably to use at their next job.  In both cases the code was for financial companies’ high-speed trading systems.

Let’s acknowledge that high-speed trading systems are more complicated than, say, a content management system.  I hope they are anyway, though recent market activity makes me doubt that sometimes.  Anyway, the algorithms are complicated, probably convoluted, hopefully rigorously tested, and they contain key competitive information that other companies would probably want to take a look at.  It’s valuable stuff.

That said, if a programmer doesn’t know how his systems work, he’s probably not worth the $1.13 million salary the new company was offering him (note to self: I’m in the wrong market, clearly.)  Let’s assume that the new company wasn’t simply out to buy a copy of a competitor’s source code – they obviously wanted someone with experience in these matters, and knew they’d be getting some insider knowledge in the bargain whether it was ever talked about or not, but I’m willing to assume they knew they were hiring an expert in a certain subject area.

Of course, if you’re an expert, you probably don’t need to copy source code.

Here’s the thing: a good programmer is always thinking about how he’d do things better if he had a chance to do it again.  A good programmer also knows how to evaluate build versus buy, and can resist the lure of not invented here system, but ultimately if you’re paying someone a premium for what’s in their head, it’s what’s in their head that you’re paying for.  Yes, that’s circular, but it’s not about what notes or code they can swipe on the way out, and if your business needs that information to stay competitive, there’s a problem.

I’m not advocating open-sourcing competitive assets here, but the reality is that, for most businesses, most of the value in keeping your code secret is to create a (mild) barrier to entry for competitors.

I realize this argument would be easier to make if the code was for something where a few less dollars were involved and the competition wasn’t so fierce, and there are probably people reading who can make a case for industrial espionage at that level, but speaking as a former developer, assuming you’re going from one company to another who both have successful systems already in place (so we’re not talking about ramp-up costs,) the value of that code (assuming it’s not what you’re actually being paid to bring over) is a lot less than you realize, and if it’s not, you need to rethink your salary expectations.