Worse is Better
Lisp: Good News, Bad News, How to Win Big (link)
Richard P. Gabriel says that systems like Unix, and C got popular because they did the easy thing, and not the right thing. Easy thing gets built fast, is easier to understand and spreads like virus. Lisp isn't popular becuase it is too good. Build a thing that 50% good, spread it like a virus then try making it 90% right. Worse is better.
Summaries:
Worse is Better is Worse (link, pdf)
A friend of Gabriel, Nickieben rebuttals the idea presented by Gabriel. There is no popular "worse is better" philosophy as Richard lets us believe. Nobody wants to build bad systems, C and Unix were better than the alternative. There were limitations imposed by the hardware. The "right" thing depends on the context, and the user. And the advice "worse is better" distorts young minds to not intentionally aim for the right thing. That is wrong.
Is Worse Really Better? (link, pdf)
Gabriel adds some more arguments to support his idea in context of OOP. Due incremental change being easier to adopt than radical improvement, due to "the right thing" being less habitable, the "worse is better languages" like C++ are taking over in OOP. While languages like Smalltalk, Eiffel and CLOS are clearly the right thing.