Slashdot said that programming has a political axis with conservative and liberal programmers. Liberal programmers worry about how fast you take to implement a feature. Liberal programmers need the answer to this question. (1)

Paul Graham said that his company could implement features faster than his competitors, because his company programmed in Common Lisp and Common Lisp makes you twice more productive as a programmer. (2)

Time passes, progress happens, Common Lisp was the fastest, a more modern programming language is the fastest, so today what programming language is the fastest to implement features? Why?

(1) http://slashdot.org/index2.pl?fhfilter=political+axis
(2) http://www.paulgraham.com/avg.html

As far as I know, pointless flame-wars are the fastest to develop: my fav linux distribution, my fav unix-variant, my fav text-editor, my fav desktop-environment, my fav programming language etc...

2 members found this post helpful.

The correct answer is: it depends on the feature.

What he said.

But in general it seems a three-way tie:
Lisp (yep, still near the top)
Smalltalk,
Forth.

Unless you like really low level grunt work (near assembler) you will avoid Forth. It is natural and fast to extend the language and add features, but you are starting from such a low level that you will not care unless you work in that space normally. If you want to play with it, look for online tutorials and a good book. The mindset is like that required for good assembly language programming.

Smalltalk was an excellent Object Oriented engine long before JAva seemed a good idea, and seems to have only gotten better with time. Alas, I do not think any of the best engines are FOSS, but the mindset is the same as that required for OO programming in Java, Perl, or Delphi.

List is still (in many ways) the best for adding features quickly. There are FOSS implementations. The mindset is totally unique in my experience, but once you GROK LISP you can do almost anything worth doing faster.

There are other fine candidates in this space, and someone with different experiences may disagree without either of us being wrong.

All of that said, there are very few times when the language that is best for a particular factor is the best for any particular large project. Everything is a compromise, and picking a language for a project or selections of projects is an excellent example. You must consider who will support and maintain the projects after completion, document eh projects, where it will be implemented, and how the language selection affects all of those and other factors as well as the obvious coding issues.

Why do you ask?

The Firefox team is taking 6 weeks to implement 3 features, but I want Firefox to go faster.

3 features in 6 weeks is actually extremely fast turnaround. Especially on a project of Firefox's size and complexity.

1 members found this post helpful.

How many bugs are you prepared to accept?

Haste makes waste

1 members found this post helpful.

I think is java or at least java is one of them because it let's even non-fully-aware programmers to make sources without so many bugs. C it's a nightmare, it's still my favourite programming language but I'm trying to stop using it. Prolog and it's spin off sound very interesting but performance it's an issue and so on

OK, disregard my earlier answer. It was totally correct but answered the question you asked instead of the question you meant to ask.

It is a totally different thing to add features to an existing package or application than it is to add features to a programming language. You need to consider all of the existing code base, what other parts interact with the part you want to change, and how your change will affect the way every other part of the program will react to your change, then code not only the change, but every modification in every other part to adjust for the differences in behavior. There are advantages to certain languages over others for this, but they are generally trivial compared to the other factors.
(I would avoid PILOT and COBOL though.)

With a small program this is fast, but takes a clear head. With a large project like Mozilla, it takes careful organization, coordination, documentation, oversight, testing, and TIME. (Time to check, double check, and when you are sure you have it right check again because you have certainly missed something.)

If you want to speed that up, go ahead. You do it by becoming an excellent programmer, then an excellent TEAM programmer, then offering to help change the code. (You will not speed it up much, but it will keep you too busy to complain while you wait. It might also help the project, and is well worth investing your time if you really care.)

RE: Java. I am very sorry. Java has a few nice features and great potential as a concept, but in practice it is not really the best at anything at all. As a compromise it is interesting, but I no longer use it. IF you code for portable devices and follow a few strict guidelines it can serve you well, I am not disrespecting it, but it is seldom the BEST answer for anything. The costs of using it, many kinds of overhead, are just higher than I can afford. If it has a BEST factor, it is very useful for teaching Object Oriented programming if you cannot use Smalltalk.