I think people who read this essay should keep one idea in mind. You cannot have the best method of doing something on the first try every time. Another words, a bunch of failures is a significant sign of a few great achievements.

Look at other languages and tell me what great features they've introduced. How about their failures. You should compare all the languages on their failures and achievements before assuming the worst of what he says. To me a language that does nothing new is a failure. Even a language that only has failures is not a complete failure, you at least get somewhere by finding out what doesn't work.

Although, I find it interesting that he has left out a couple of things that have affected programmers the most in the last 20 years.

1. Marketing and corporate backing. Which don't seem even to be remotely related to quality or success of a languages features.
2. Libraries. To me, the true health of a language is dependant on the health of it's libraries. What good are small sets of axioms without enormous libraries for doing everything. Instead you are left with a language that is the equivalent of a doorstop...but worse, you have to write 1000 lines of code to make it hold the door open. This saying holds well for perl (it's someones sig), 90% of every perl program is already written (not sure if quoted correctly). How many other languages can you say this about.

Finally, I have a genuine question to ask. I don't memorize many popular people...hack or otherwise. Exactly how much perl does Paul Graham know?

smiles