3 ms·

The whole Unix "worse is better" thing seems to me like a specific instance of a more general principle. You see it in network effects and first mover advantage

The whole Unix "worse is better" thing seems to me like a specific instance of a more general principle. You see it in network effects and first mover advantage in business, but I saw it in other industries I worked in as well. When I was a military officer, it was drilled into us to move as soon as we developed 80% of a plan, in part because you can make up the rest as you're going but also because just striking first is itself a tremendous advantage. Being faster than the enemy is often better than being better than the enemy. In finance, we were pushed to trade on imperfect strategies because any edge whatsoever disappears almost instantly as soon as you use it and you have to assume someone else is going to figure out whatever you figured out. I don't think you can generalize to just first mover always win, especially since Unix wasn't the first mover. More that any plan, scheme, product, whatever that can actually work needs to be put into operation as soon as it can work. It's no guarantee you won't lose anyway, but you can win for a very long time until you don't. Unix may be supplanted at some point and Bitcoin too, but Tom from MySpace at least made a few hundred million when he likely makes nothing if he launches after Facebook.

chriswarbo 6y ago
I wouldn't use "Worse is Better" as a stand-in for first-mover advantage, since (as you say) there are times when the benefit of speed more than pays for the increase risk of unknowns. I'd reserve Worse is Better for cases where (a) that speed advantage was gained by specifically choosing to ignore/hack-around a known problem, and (b) that problem/hack becomes worse due to the inertia of this approach. The classic example is C, which has spread far and wide, as a foundation for almost all of the modern computing house-of-cards. Yet this adoption has come with memory management problems (and their associated security exploits), a foot-gun approach to macros, etc. when such issues are literally non-existent in contemporary languages like Lisp, Scheme, ML, Prolog, etc. A military example might be the trench warfare of WW1: a quick way to avoid machine gun fire, which froze armies in place for years. A financial example might be loan sharks. In Bitcoin's case, the hack of trusting the longest chain and incentivising hashcash to reduce spam/overtaking has resulted in today's energy-hungry monstrosity.
ska 6y ago
> The whole Unix "worse is better" thing seems to me like a specific instance of a more general principle. That was the contention of the original article, if I recall it correctly.

chriswarbo 6y ago

I wouldn't use "Worse is Better" as a stand-in for first-mover advantage, since (as you say) there are times when the benefit of speed more than pays for the increase risk of unknowns. I'd reserve Worse is Better for cases where (a) that speed advantage was gained by specifically choosing to ignore/hack-around a known problem, and (b) that problem/hack becomes worse due to the inertia of this approach. The classic example is C, which has spread far and wide, as a foundation for almost all of the modern computing house-of-cards. Yet this adoption has come with memory management problems (and their associated security exploits), a foot-gun approach to macros, etc. when such issues are literally non-existent in contemporary languages like Lisp, Scheme, ML, Prolog, etc. A military example might be the trench warfare of WW1: a quick way to avoid machine gun fire, which froze armies in place for years. A financial example might be loan sharks. In Bitcoin's case, the hack of trusting the longest chain and incentivising hashcash to reduce spam/overtaking has resulted in today's energy-hungry monstrosity.

ska 6y ago

> The whole Unix "worse is better" thing seems to me like a specific instance of a more general principle. That was the contention of the original article, if I recall it correctly.