Worse is better?

'Worse Is Better' is a bit of programmer lore that has fairly wide acceptance.  People often accept this as the reason why VHS triumphed over Betamax, and why people use Microsoft products.  The standard essay about it (see this and this) is about the problems Lisp has had becoming widely used.

I was reminded of this hearing that SGI, which got its start making high-end graphics workstations, has gone bankrupt.  I'm not sure I think 'worse is better' is ever an explanation of one technology winning over another.  In most cases it isn't so much that one is better than another that is important, it is how broadly useful the technology is.  We see this over and over.  The IBM PC was more popular than Apple because it was a more flexible architecture, even though, for what it did, the Apple was at least as good.  VHS had a longer recording time than Betamax and was more widely licensed, so it was more generally useful.

The explanation for this sort of thing is more along the lines of Clayton Christensen's Innovator's Dilemma than 'worse is better'.  The dilemma is that disruptive technologies emerge that address the low end of your business.  This is fine at first, since there is little margin at the low end.  Unfortunately these technologies tend to improve to the point where they take more and more of your business until yours is gone.  All during this process, people are using the 'best' technology for their purposes.  I suspect this happened to SGI's original graphics business as standard PC graphics progressed from terrible to amazing.  You can see a very similar process going on as more and more computers are based on the x86 chip architecture rather than special-purpose designs.

--Th