I don't disagree with his portrayal of the brilliant failure, or even with the statement that such a person would be attracted to Lisp. But brilliant successes and brilliant failures both seem to be attracted to it, which implies that the brilliance is responsible for the attraction, not brilliant failure specifically.
He's even further off when he says that the reason Lisp isn't more popular is because of the mysterious quality that attracts such people. It's unpopular for more mundane reasons: the syntax looks odd, and there's no good, standard implementation with lots of libraries.
"...there's no good, standard implementation with lots of libraries."
I think his point was that there's no good, standard implementation with lots of libraries because many of the people drawn to Lisp have big ideas but are bad at finishing things.
The problem is that the main dialects of Lisp (CL and Scheme) date from the era when languages were specs. Now they're implementations. It just happens that no one has released a good new Lisp dialect in that sense of language. But no one (as far as I know) has released a good new Fortran in that sense either.
Language designers who have what it takes are quite rare: Matz, Guido van Rossum, Larry Wall. It just happens none of them chose to create a Lisp dialect. But the problem is not the brilliant failures who are drawn to Lisp, but that e.g. Matz wasn't (quite).
I think you're right about those mundane reasons, but the community can be a major turnoff too. Try posting a stupid question[0] on comp.lang.lisp - you'll almost certainly get flamed, and maybe get harmful responses[1]. Try the same on comp.lang.ruby or comp.lang.python and you might get some gentle correction, but not outright hostility.
[0] A stupid question would be anything you can find out by searching google, reading the manual or possibly digging through a few thousand lines of source code.
Example [1] was just an individual nutter. It gets used over and over to show that Lisp newsgroups are mean. One sign they're not is that people have to keep re-using that one example instead of finding the new ones that should in theory be very common. From what I've seen, people on Lisp newsgroups are much like on any other.
I just went looking for recent examples to prove my point and found none. Perhaps my memory of CLL is inaccurate, or it has become more friendly recently.
Comments
I don't disagree with his portrayal of the brilliant failure, or even with the statement that such a person would be attracted to Lisp. But brilliant successes and brilliant failures both seem to be attracted to it, which implies that the brilliance is responsible for the attraction, not brilliant failure specifically.
He's even further off when he says that the reason Lisp isn't more popular is because of the mysterious quality that attracts such people. It's unpopular for more mundane reasons: the syntax looks odd, and there's no good, standard implementation with lots of libraries.
"...there's no good, standard implementation with lots of libraries."
I think his point was that there's no good, standard implementation with lots of libraries because many of the people drawn to Lisp have big ideas but are bad at finishing things.
The problem is that the main dialects of Lisp (CL and Scheme) date from the era when languages were specs. Now they're implementations. It just happens that no one has released a good new Lisp dialect in that sense of language. But no one (as far as I know) has released a good new Fortran in that sense either.
Language designers who have what it takes are quite rare: Matz, Guido van Rossum, Larry Wall. It just happens none of them chose to create a Lisp dialect. But the problem is not the brilliant failures who are drawn to Lisp, but that e.g. Matz wasn't (quite).
Are you saying that if someone releases a good new Lisp dialect, then it will become popular?
I think you're right about those mundane reasons, but the community can be a major turnoff too. Try posting a stupid question[0] on comp.lang.lisp - you'll almost certainly get flamed, and maybe get harmful responses[1]. Try the same on comp.lang.ruby or comp.lang.python and you might get some gentle correction, but not outright hostility.
[0] A stupid question would be anything you can find out by searching google, reading the manual or possibly digging through a few thousand lines of source code.
[1] http://groups.google.com/group/comp.lang.lisp/msg/4c95766a8b...
Example [1] was just an individual nutter. It gets used over and over to show that Lisp newsgroups are mean. One sign they're not is that people have to keep re-using that one example instead of finding the new ones that should in theory be very common. From what I've seen, people on Lisp newsgroups are much like on any other.
I just went looking for recent examples to prove my point and found none. Perhaps my memory of CLL is inaccurate, or it has become more friendly recently.