Here’s a gem from former Microsoft CTO Nathan Myhrvold: “Software sucks because users demand it to.” [“Why Software is So Bad”, July/August Technology Review.]
The article is a good read. But the ultimate conclusion is troubling: the author suggests that bad software is systemic in the industry, and that regulation is sure to follow. (To his credit, he cites numerous examples of easily fixable software bugs that had catastrophic results. He’s right: there’s lots of bad software out there.) But product liability law as the savior? “[W]hen the costs of litigation go up enough, companies will be motivated to bulletproof their code.”
References to the liability paradigms installed to handle the railroads, automobiles, etc. are persuasive, but is this really the best way to improve software quality? A couple things are certain: software is increasingly pervasive. Software developers consistently incorporate new features, enhance existing functionality. This ubiquity means that entire businesses are dependent on intangible lines of 1s and 0s. If those binary digits fail, does the company have any recourse?
What worries me is that if this is where we’re headed, software companies won’t necessarily try to make the software bug -free: they’ll just make it do less. But less exposure for the developer means less innovation for the industry. In this economy, that can’t be a good thing. (I’ll admit to a bias: I’m part of the management team at a software company. But I’m also a software consumer – and I’m not sure I want or am entitled to a dollar every time I see a blue screen of death. Well, come to think of it…)