Moldable Development in CLOG - Konrad Hinson (from Europe, (in the form of a screencast))

open discussions

“why Common Lisp instead of Scheme or Racket?”

“do you have an opinion on Common Lisp vs. Scheme (Racket)?”

`1. Historical. I had adopted Common Lisp as my main Lisp before even thinking about this project, so CLOG played no role in the decision. I had been using Racket intensively for a few years, and then discovered Smalltalk, which I have also used intensively for a few years. I had learned to appreciate the language malleability of Racket and the interactivity and introspection of Smalltalk. So I ended up adopting Common Lisp as the middle ground: it has both of these features in imperfect but usable form.

`2. Technical. Most Schemes (never say “all” with Schemes!) don’t have the required level of introspection to write inspectors such as this one. I’d love to have something similar for Guile (which I use a lot because of Guix), but that would be complicated and perhaps even impossible because Guile doesn’t let you extract much information about an object in memory. Racket may have what it takes, though I am not 100% sure.

`3. Social. Common Lisp is boring technology: widely used across application domains and platforms, well understood, warts included. There are decent to good libraries for many things I want to use but not make my specialty. No Scheme implementation comes close to that. Racket probably has the best library support, but it still pales compared to Common Lisp. Platform support is another point: I can use Common Lisp on all my computing devices (including in particular my Android e-ink tablet).