How did you find newLISP?

[lithper]

Thanks, that was interesting to read.

I found newLisp very recently, but cannot remember where. I was going through a number of implementations, and when found newLisp, was really struck by how well thought-out the package was, that I could do scripting immediately, that documentation is informal, modern (full of examples and cookbook recipies, rathern than the insane formality from 1970s with "big" lisps).
All APIs were there, the tiny standalone executable could be started in tens of copies or used to distribute scripts between machines - I was fascinated by how well thought-out it was, as I already said. A true hacker's tool in the original sense of this word.

ow to make it more popular?
I try to mention newLisp where I can, but cannot say I succeeded much so far.

One way could be to write (why do not we pool our efforts together here?) some article under a good catchy title and publish it in some Linux on-line magazine, the Perl Journal or similar (under the guise of comparison with the perl approach or sth) , push NewLisp as "perl 6 that is already here" ( that is the way I think of NewLisp myself, in my head - this is much of what I would dream perl6 to be - but for the problem that perl6 is NOT except as a fantasy, and it's completely unknown when it could get born into the world )

I'd try to find a well-known sysadmin, linux and/or scripting language on-line publication and try to write a series or a column there.

..maybe also write some "newLisp poetry" - like that Perl "poetry", when the text of the program that parses OK reads like a poem ;))

[Tim Johnson]

In 2004, I posted a question to a scheme mailing list, said a little about
my needs and solicited comments for a recommended scheme.
An individual on this list, who also codes in CL, scheme and prolog
mentioned newlisp. It has been just in last few months tho', that I
have been seriously learning newlisp.

I currently also code in Python, rebol, and javascript. After many years
coding in C/C++, I pretty much retired that codeset in 2001.

[hsmyers]

I got here from a URL on Hacker News. I've been quite happy ever since. And while I'm implementing my current project(porting my CPAN PGN: modules) in both newLisp and common lisp(clisp via slime and emacs) and will eventually get around to scheme, ruby and python, I've got to tell you that newLisp wins hands down. The multiple languages help me make improvements across the board and is an excellent way to keep familiar with any changes there might be.

As an longtime autolisp programmer I was always searching for lisp-alternatives running outside autocad but fairly compatibel with autolisp.
At that time (around 2001/2002) I evaluated every lisp I get my fingers on. (xlisp,corman-lisp,Lispworks etc.)

HPW, do you remember in your evaluations how much xlisp resembled autolisp— just wondering how much divergence there has been since Walker borrowed xlisp from Betz(sp?)?

–hsm
p.s. I started with AutoCad 1.4, haven't done much since 200i...

[Tim Johnson]

I programmed autolisp a bit in 1993-1994. An experience that I think opened my mind to newlisp.

I _do_ believe that python might be a better choice for a team programming environment where there are multiple programmers working on the same codebase. I attribute that opinion to python's stricter engineering OOP approach and scoping. It isn't an issue for me, however, as in my company, my partner and I work on different code bases - he's a perl monger...

Given that the rebol developer's are indifferent to 64-bit issues and rebol is not open source - it is likely that where I would be otherwise working in rebol, I will soon be working in newlisp.

[Jeff]

Python is nice for many things, but I prototype even faster in newLISP. I use Python for many of our main applications, but I am finding that in many cases, I end up using newLISP or common lisp due to their speed.

[Tim Johnson]

Jeff, - just curious - do you have more than one programmer working on one codebase?

I find development much faster in rebol than in python, and that includes
prototyping, and I would imagine similar comparisons between
newlisp and python.

When the customer expresses the concern that - should I be no longer
available to maintain code or should he/she have to find another
programmer for any reason at all - I have to concur that it would be
easier to find a python programmer than it would be to find a rebol
programmer.

In the case of newlisp, I'm sure that any scheme or CL programmer worth
their salt would be able to take over newlisp code. Especially if the
programmer documents well and documentation for both newlisp itself
and the custom code is part of the code package delivered.

If I'm right, it is a selling point that rebol doesn't have....

[Jeff]

I can't speak for rebol, but yes, with multiple programmers, Python is easier to maintain. The cost of that is that you lose a lot of flexibility and speed.

I love Python. I use it for most projects. But when speed is a real issue, I use cl, which can be compiled. I have used ocaml before as well, but it does not have the packaging tools that cl has (like asdf) or support for threading.

newLISP, though, is what I often use initially when I need to speed something from python up. It's much faster than Python, but few interpreted languages can keep up with compiled code.

I like newLISP because I can concisely write a huge amount of functionality into a simple, small, space (the same reason I like Python). And in many respects, it is easier to maintain, because it's a bottom-up language.

For example, I am now rewriting a program I wrote in common lisp in newLISP (which means I am having to write a *ton* of support code that was already written for me in cl.) However, it does not get a huge boost from compiling. What it really needs is high concurrency, responsiveness, and low memory usage, which is what nl offers me.[/i]

[Tim Johnson]

Although the syntax is different, there are a lot of similarities
between rebol and newlisp. I like the way that emacs handles
newlisp - it "understands" newlisp better than it does rebol.
And there is some improvement in the way that newlisp reports
errors (over rebol), but python really nails errors right down to the
file name and line number. Can't beat that.

[Jeff]

Python does have much better debugging tools than almost any other language (note that I did not say error handling; I prefer conditions to exceptions.)

One thing that I **really** miss in newLISP is a profiler. I might have to write one using the new hooks into the interpreter.

[aron]

I learned Lisp at Linköpings university in the introduction programming course, we did use Allegro Common Lisp(Franz, Inc.). Later I think I found newLisp while locking for free implementations of Lisp, also working for windows XP, because that was what I was using at the moment.

I don’t remember exactly but I think it was from this list I found it:
http://wiki.alu.org/Implementation

[pavelludiq]

This morning I found the site and discovered that i can write this:

Code: Select all

('(1 2 3)  1)

and get 2, oh simplicity!

I also found out that i don't have to run newLISP on top of java(like clojure) or PLT(like arc), and i still get a modern dialect of lisp with an active community and descent libraries.