Why You Do Not Use Lisp?
|I use Lisp, at least sometimes.||49 (44%)|
|It has not enough support for modern programming practice (libraries, threads and similar.)||42 (37%)|
|Its syntax repulses me.||26 (23%)|
|It is socially inadequate (small, not very alive community.)||23 (20%)|
|It doesn't have any technical advantage I might want.||20 (18%)|
|Its IDE is not good enough.||18 (16%)|
|I'd like to use Lisp, but my superiors or colleagues do not want that.||15 (13%)|
|It has not enough support for higher level programming (laziness, logic programming...)||15 (13%)|
|It is socially inadequate (arrogant, unhelping community.)||14 (12%)|
|It is too slow or bloated.||12 (10%)|
|It has attractive technical advantages, but these are hard to learn, use and not worth effort.||11 (9%)|
|It is OK, but only in some commercial version - and I want free or much cheaper one.||10 (9%)|
|It is so marginal that I never considered it.||9 (8%)|
Presented on this way, it is not easy to recognize the message of the poll. All answers are expected, but what their relative importance means? However, I believe we can filter out important information, if we take into account two things: first, voters have much higher than average interest in Lisp and functional or metaprogramming than usually. Those who do not care for Lisp didn't care to vote on this poll as well. About 13% of voters would like to use Lisp, but they cannot because other their colleagues and bosses do not allow it. This is surprisingly low number of people - whatever your favourite language, the chance that your boss will prefer something else is large. If anything can be concluded it is that majority of Lispers somehow managed to group together or work on their own.
Lot of voters (37%) miss some libraries, or support for some modern feature. However, it doesn't say us much: lack of the libraries is universal phenomenon. I believe many people would say they do not use some mainstream language, say, Python because they've found that Ruby, Java or Perl have some specific library supported better. Second, this reason is somehow circular - less libraries means less users etc. Taking it into account, even if Lisp community wrote libraries as no one did, the result wouldn't be much better. It would be much better regarding libraries, but not much regarding programmers. Many of these who complain about lack of libraries probably already use Lisp part of the time.
The second reason is much stronger: 23% of voters do not like Lisp syntax. I incline to think that very few people who do not like Lisp syntax use Lisp on the first place. So, almost the half of the voters who do not use Lisp (56%) - on this poll - do that because of the syntax. If we assume that interest of the voters for Lisp is above average, we can safely say that Lisp syntax is the most important reason for majority of people who do not use Lisp.
One surprise, for me, is popularity of the answers "Lisp doesn't have any technical advantage I may want." (18%) "it has not enough support for higher level programming" (and it includes static typing) (13%) and "It has attractive technical advantages, but these are hard to learn, use and not worth effort" (9%) we have another revealing information - Lisp community didn't succeeded to develop or demonstrate technical advantages of Lisp. Functional languages programmers certainly understand these advantages, but they still think these are not that significant.
Good news is that Lisp got rid of its image as bloated and slow. Only 10% of voters complained about that. For, say, Modula-2, even that would be too much, but for Lisp family, which is certainly not designed to be among the leanest possible languages, the result is quite good.
Social inadequacy, i.e. small, not very alive community - rates surprisingly high (20%). I had no clue that people like to communicate that much. One might think that large community means lot of libraries. It does - but it still doesn't mean that voter will get the library he wants. So, it is, I believe, more of a human instinct for joining large, progressing, vibrant, active community, perhaps related to survival. Well, good information for all Lisp programmers and especially my Newlisp fellows; communication matters. On a related matter, significant number of voters (12%) believe that Lisp users are arrogant and unhelpful community.
Some 16% of voters missed good IDE. Well, two commercial Common Lisps have good IDE, and have free personal versions. PLT Scheme has nice IDE as well. Newlisp has very simple IDE. Personally I use Scite, which is text editor that can cooperate with command line tools easily, and it is good enough for me. About 9% of voters miss something in free versions, and it can be easily that it is again IDE.
Another poll that might interest you:
Opinions on Eval in Lisp, Python and Ruby. The Results of The Poll.