Something that would be of more interest would be a comparison of Scheme and PicoLisp. I think even people who who use CL a lot are kind of embarrassed about a lot of the design choices, and recognize that it was meant to be a practical hodgepodge of features from different lisps, rather than a focused design philosophy.
Of course, a lot of the same points below would still apply. But with Schemes the whole issue of composability vs monolithism as mentioned in the chrisdone.com article pretty much melts away. SRFI-1 has all the "haskell" functions mentioned in the article. I work with Guile Scheme, and to me the biggest differences between Guile and PicoLisp are: (1) Guile code is compiled (though often at runtime); and (2) Guile is designed from the ground up to interface well with system libraries (you can even embed Guile inside a C program fairly easily). Of course, scheme is really just a language specification and design philosophy, and implementations get even more fragmented than CL does, because Scheme is more minimalistic and extensible than CL. So, perhaps a comparison of Scheme and PicoLisp would be even more challenging. On Tue, 2017-09-26 at 17:07 +0200, andr...@itship.ch wrote: > Disclaimer: I don't really know Common Lisp. > > Just a try, use it as inspiration not as answer, would need re- > wording to be used as quora answer: > > - they're about the same age, both inspired by Maclisp, both used > commercially since the 1980s > - picolisp is a language and runtime VM - CL is a language > specification with multiple implementations > - CL libraries may often be platform-dependent (or depend on a > certain compiler to be used) - code in picolisp has much less such > dependencies > - picolisp has no compiler, purely interpreted by design - CL is > usually compiled, interpreted only during development > - picolisp directly evals s-expressions, the structure of the memory > representation is the same as the structure of the source code - > common lisp compiles to XX (?) > - only 3 strong types: list, number (arterially big signed integers), > symbol (with property list), all other types are dynamically duck- > typed - common lisp has many types (?) > - picolisp uses F-Expressions (FEXPRs, Functions which decide > themselves if/when/how to evaluate arguments) - common lisps uses > macros (rewrite code). F-Expr are slightly more powerful but cannot > be compiled > - common lisp as big community - picolisp community is pico > - many libraries for CL - in picolisp programmers usually re-use > libraries from other languages > - picolisp strongly follows unix philosophy - CL does not follow unix > philosophy ( http://chrisdone.com/posts/haskell-lisp-philosophy-diffe > rence ) > - picolisp has integrated graph database, persistent database objects > are first-class citizens, integrated prolog engine > - picolisp has no multithreading, instead picolisp has integrated IPC > to work with multiple processes and makes asynchronious programming > easy - common lisp has ... ??? > > Please correct my errors. > > Most points of the FAQ should probably go into the quora answer: http > ://www.software-lab.de/doc/faq.html > See also: http://picolisp.wikidot.com/ > > > > ----- Original Message ----- > From: Mike Pechkin [mailto:mike.pech...@gmail.com] > To: email@example.com > Sent: Tue, 26 Sep 2017 15:58:05 +0300 > Subject: Re: PicoLisp on quora.com > > > > > > > Uh, oh, this is really tough. > > > Just one smart paragraph. > > (mike) > > > -- Christopher Howard Enterprise Solutions Manager Alaska Satellite Internet 3239 La Ree Way Fairbanks, Alaska 99709 1-888-396-5623 https://alaskasatelliteinternet.com personal web site: https://qlfiles.net https://emailselfdefense.fsf.org/en/
Description: This is a digitally signed message part