Laboratory Robotics with Common Lisp
To stave off the never-ending questions -- variations on the theme of "Where is Loper? Why the wait?" I would like to confess the following: I live a double life!
I spend my days... working! For money! So that I can eat.
Here is my current commercial project.
It is a laboratory robotics controller, based on the Programming by Demonstration paradigm. The beast is made of ~10,000 lines of Common Lisp, sitting on SBCL/Linux. Much of the code relates to semi-automated protocol reverse-engineering (for adding support for new lab instruments.)
That's cool!
And then you save money! So that you can quit.
Dear Stanislav, you might enjoy the little Lisp book at
http://www.civilized.com/files/lispbook.pdf
Dear gary knott,
I have a well-thumbed printout of your book in a three-ring binder! It was one of my early (high school) reads on the mechanics of simple Lisps. Thank you for writing it!
Yours,
-Stanislav
gary: thank you for posting this. A very enlightening read.
Really cool! Could I ask why your company chooses common-lisp instead of C/C++ to implement the robot controller?
Dear jie,
Lisp (or something equivalent) is the only rational choice for an exploratory engineering project with a small number of non-interchangeable developers.
Think of situations which call for craftsmanship rather than industry; where only the results (rather than process) matter. You will find Lisp there. Or rather, you probably won't, because most such projects spend their whole life cycle behind firmly-closed doors.
See also: this.
Yours,
-Stanislav