Modern, functional Common Lisp: myths and best practices

Re:Modern, functional Common Lisp: myths and best practices by lispm on Tuesday February 25, 2020 @12:47PM (#22415293)

Yes, generally any source to byte-code translator is a compiler. The byte-code then can either be interpreted or compiled, too. Since the default CPython implementation interpreted the byte code, it was said it's an interpreter - but strictly speaking it is not.

Originally interpreters were implementations which execute source code - which in Lisp is widely available. For example SBCL for a long time only had a native code AOT compiler, but now also includes s-expressions.

Famous is also the Lisp interpreter in Lisp from McCarthy, where he defined the core language in itself. See the paper from Paul Graham about that: http://languagelog.ldc.upenn.edu/myl/llog/jmc.pdf That's a very primitive Lisp interpreter - many real ones are implemented in C, Assembler...

Just be aware, that in the Lisp world 'Interpreter' means something very specific: an interpreter of the source language.

[ Parent | on HN ]

Re:Modern, functional Common Lisp: myths and best practices by bibyte on Tuesday February 25, 2020 @12:58PM (#22415410)

Fair enough. My assumption was that you would consider a source to bytecode compiler an interpreter (to be fair most people do). The next time I say something about Common Lisp I will list all the implementations instead of saying something as simplistic as "compiler" or "interpreter".

[ Parent | on HN ]

Re:Modern, functional Common Lisp: myths and best practices by lispm on Tuesday February 25, 2020 @01:11PM (#22415555)

It's also complicated because some implementations support several modes. ECL for example has a Lisp interpreter, a byte-code compiler&interpreter and a native compilation via C. All code variants can work together and call each other.

[ Parent | on HN ]