← Back to context

Comment by AdmiralAsshat

2 years ago

There's something to be said for throwing everything out and starting from scratch every once in awhile. How many decades old is Unix now? How many is Windows, even?

You lose the years of battle tested code, sure, but you'd also lose decades of baggage, tech debt, obsolete code for things no one has had to worry about for 20 years, etc.

From the perspective of the company, you also have a moat in the form of a proprietary system that doesn't have to compete with anything else because little software your customers depend on will run on anything else. Your business decisions become the Word Of God to people and companies with no real alternative.

BeOS never got successful enough to do that, but A Particularly Loud stock would argue this can work fairly well.

  • BeOS followed Apple's incompatible walled garden philosophy but it came out at the same time as the Web, Netscape, Java, Flash, RealPlayer, etc. Suddenly portability and interoperability were huge and BeOS was totally unprepared for it.

I think you underestimate the amount of functionality in something like the Linux kernel and how much work it would take to recreate even a small percentage of it.

  • I was a huge BeOS fan at the time, but in retrospect it was more like a vertical slice demo of an operating system. There was so much missing features and technical debt that fleshing it out would have been 90% of the work.