Much of this looks good; however, I really hope that "unfriendly" stuff isn't taken to heart. (Or maybe everyone can take it to heart, except for one language community, and then that becomes the "acceptable Lisp".)
On the other hand, I'm interested in a different set of issues. Consider Haskell. As a language, it's not doing too badly. Why? Well, one reason is that it is very well designed. But another is Hugs. Hugs was available a decade ago, it ran anywhere, was lightweight, let you use your favorite editor, implemented almost the complete standard, let you write and run Haskell code easily immediately after installation (which was fast), and, while it's a bit short on tools, it does let you write and run "real" programs.
I would say that Lisp desperately needs a Hugs.
If I look into getting a Lisp implementation, I see that they want me to learn Emacs, or mess with the JVM, or are not available on all platforms, etc. None of these things have much to do with the language itself, and they are all barriers to adoption.
I just want to click a box in my package manager, hit the button to download, run the program, and be in a REPL, or maybe say "edit" (or whatever) and be typing in a source file, which I can easily compile and execute.
Again, Lisp needs a Hugs. But Hugs was designed years ago, and things have changed. What does "a Hugs" for today look like? I'm not completely sure, but the above is a start toward describing it, I think.
Hehe, you are telling my hearts content. I think something along those lines need to be thwapped into the face of every designer and implementer of a new language:
- do not take away my vim for your puny IDE. I rather won't try your language than learn a new editor
- have a reasonable compiler out there which follows the guidelines to a compiler/interpreter interface on the command line: either "myLanguageCompiler program; ./program" or "myLanguageInterpreter program".
- get this reasonable standard compiler into the major linux repositories
Basically, all of these boil down to a single important point: Lower the entry barrier to your language considerabely, even to a point where things are just reasonable and not great for the new programmer.
Certainly, vim might not be good for editing (say) Lisp, and the standard compiler called (say) lisp or lispc might not be fast and efficient for everything, but I don't care about that if I just want to spend an evening toying around with the language.
On the other hand, I'm interested in a different set of issues. Consider Haskell. As a language, it's not doing too badly. Why? Well, one reason is that it is very well designed. But another is Hugs. Hugs was available a decade ago, it ran anywhere, was lightweight, let you use your favorite editor, implemented almost the complete standard, let you write and run Haskell code easily immediately after installation (which was fast), and, while it's a bit short on tools, it does let you write and run "real" programs.
I would say that Lisp desperately needs a Hugs.
If I look into getting a Lisp implementation, I see that they want me to learn Emacs, or mess with the JVM, or are not available on all platforms, etc. None of these things have much to do with the language itself, and they are all barriers to adoption.
I just want to click a box in my package manager, hit the button to download, run the program, and be in a REPL, or maybe say "edit" (or whatever) and be typing in a source file, which I can easily compile and execute.
Again, Lisp needs a Hugs. But Hugs was designed years ago, and things have changed. What does "a Hugs" for today look like? I'm not completely sure, but the above is a start toward describing it, I think.