Name: Anonymous 2013-08-11 20:42
After reading some of PG's essays, I did give Common Lisp the old college try, but without much success. I got the Touretzky book and did all the exercises, but didn't feel like I was learning anything useful.
The main problem for Lisp 's adoption I think is Common Lisp. You have to learn, I don't know, 4 different types of assignment, 6 different equality functions, 20 different loop constructs that were maybe revolutionary in 1965 mainframes but today look like, well, like they were revolutionary in 1965 mainframes.
Oh, and CADADR, CADADDADDR and CADADADADADDR -- is that some kind of joke? The dot... Symbol attributes (I think they were called, those useless bits of info that you "get" from a symbol). The FORMAT windings for printing a newline... The associative lists... a O(n) hashmap? again, not very productivity-enhancing in my opinion.
And don't get me started with the "don't worry about the parens, we will fix it with indentation" nonsense. No you don't fix it with indentation. I am not getting any productivity boost if I have to poke around with a ruler and square to see where the conditional ends and the else clause begins. Hint: a paren-matching editor is of no use when you are reading a printout of a program.
(They tell me Franz has a n if* macro that has "then" and "else" to make it more readable. Of course, the hardcore Lispers despise it.)
I think the best idea if you want to change the world is: rescue the useful parts from under all that pile of revolutionary 1971 technology, and I assume there must be, take the good bits (macros, continuations, clothures, whatever) and build them into a language that can be described it in a K&R-sized book and that humans can learn without having to undergo 7 years of self-Pavlovization-cum-Stockholm Syndrome. As things stand today, I'm sure macros and closures are wonderful, but there is just too much useless junk you have to memorize before being able to use them.
This is what I think Java got right: take the main object-oriented concepts from Smalltalk and leave off the 1975 conceptual breakthroughs like "the while loop is an object too", that in the end of the day are of little practical use.
Still waiting for that language that has closures and macros and that I can learn without having to take a year of unpaid leave,
The main problem for Lisp 's adoption I think is Common Lisp. You have to learn, I don't know, 4 different types of assignment, 6 different equality functions, 20 different loop constructs that were maybe revolutionary in 1965 mainframes but today look like, well, like they were revolutionary in 1965 mainframes.
Oh, and CADADR, CADADDADDR and CADADADADADDR -- is that some kind of joke? The dot... Symbol attributes (I think they were called, those useless bits of info that you "get" from a symbol). The FORMAT windings for printing a newline... The associative lists... a O(n) hashmap? again, not very productivity-enhancing in my opinion.
And don't get me started with the "don't worry about the parens, we will fix it with indentation" nonsense. No you don't fix it with indentation. I am not getting any productivity boost if I have to poke around with a ruler and square to see where the conditional ends and the else clause begins. Hint: a paren-matching editor is of no use when you are reading a printout of a program.
(They tell me Franz has a n if* macro that has "then" and "else" to make it more readable. Of course, the hardcore Lispers despise it.)
I think the best idea if you want to change the world is: rescue the useful parts from under all that pile of revolutionary 1971 technology, and I assume there must be, take the good bits (macros, continuations, clothures, whatever) and build them into a language that can be described it in a K&R-sized book and that humans can learn without having to undergo 7 years of self-Pavlovization-cum-Stockholm Syndrome. As things stand today, I'm sure macros and closures are wonderful, but there is just too much useless junk you have to memorize before being able to use them.
This is what I think Java got right: take the main object-oriented concepts from Smalltalk and leave off the 1975 conceptual breakthroughs like "the while loop is an object too", that in the end of the day are of little practical use.
Still waiting for that language that has closures and macros and that I can learn without having to take a year of unpaid leave,