[pro] Learning Lisp the Bump Free Way

Vladimir Sedach vsedach at gmail.com
Fri Jan 21 04:33:58 UTC 2011


I really didn't appreciate this whole discussion, but I think I can
contribute a few points about how to think about the problem so that
things like this will stop happening.

There are two assertions being made here:

* "Lisp is falling behind"
* "To stop Lisp falling behind, we need to do <X>"

No one has provided any quantitative evidence for either.

How is Lisp "falling behind?" Compared to what? And what effect does
that have? What effect does that have on you personally?

Why do you think doing <X> will be good? Did anyone ask for <X>? Is it
actually preventing people from doing things? How can you tell <X>
will work when you can't even quantify the "falling behind" thing <X>
is supposed to change?

What is the basic pattern behind this argument? Alexander, I will use
your original post as an example, please don't take it personally:
"Lisp is dying: some JavaScript implementation is slower than Clozure;
let's work on the Clozure IDE!" Where is the logic here? Most of these
arguments use "Lisp is dying" handwaving to promote a personal
project, and in most of these cases no one outside the current Lisp
community cares about the issue.

What does it mean to say "Lisp is falling behind?" There aren't enough
people using it? Have you tried listening to people who have tried and
failed to use Lisp? Do you understand the difference between a reason
and a rationalization?

There are people who rationalize not using Lisp by saying it has
parentheses, and there are people who have a reason for not using Lisp
because they can't find the libraries they need. Listening to the
former is a waste of time, listening to the latter will give you a way
to actionably decrease the "Lisp falling behind" and measure it.

Here are some things I notice people actually saying:

1. The #1 problem is that people (and this includes even people who
have been using Lisp for decades and post to this list, not going to
name names :)) have misconceptions about what current Lisp
implementations, libraries and tools can do. And this is just the
people that know about the existence of Lisp. Simply letting people
know about Lisp and how great it is is the most direct way you can get
more people to use it.

2. There are no libraries to do <X>/I can't find any libraries.
Improve and promote the existing libraries. Write new ones for some
service/software/function that isn't possible from Lisp yet. Put a
description of your library on CLiki, ask Zach to add it to Quicklisp.
Help improve the existing library listings in CLiki. Help with
Quicklisp, add metadata and search stuff to Quicklisp projects
(https://github.com/quicklisp/quicklisp-projects). Give Zach money.
Basically anything that makes Quicklisp better is good.

3. How do I deploy applications? Web apps (the story here is
incredibly bad -
http://stackoverflow.com/questions/2458805/nohup-sbcl-ubuntu-couldnt-read-from-standard-input/2459674#2459674)?
Work on cl-launch, lispx-proxy, cl-daemonize

This is today's list, it will change tomorrow. The important thing is
to understand what's really going on, and the psychology behind it.
Which is why the last thing I'm going to say is that saying "Lisp is
dying" is destructive. If Lisp experts are saying our community is a
loser, why would anyone want anything to do with Lisp? Stop doing it.

------

Now I'm going to reply to Dan.

I think ITA's attitude of "I wish we didn't write qpx in Lisp but it's
the only thing good enough" is bizarre. That no one was assigned even
part time to be a "Lisp evangelist" was a *strategic* mistake. It
would have saved ITA hundreds of thousands in recruitment costs, and
improved productivity through an improved Free Software Lisp
ecosystem.

Another thing is the "I won't be allowed to use Common Lisp" attitude.
This is the perfect opportunity to force Common Lisp into Google, and
you're walking in expecting to fail.

Vladimir




More information about the pro mailing list