From: Christian Neukirchen Date: 2006-12-20T04:39:25+09:00 Subject: Re: Ruby for Kids? znmeb@cesmail.net writes: > Quoting Christian Neukirchen : >> >> I didn't intent to question LISP 1.5, but the significance it made. >> Say, compared to Scheme, LISP 1.5 is not that world-changing, is it? > In the context of *Lisp*, Scheme is absolutely earth-shaking. But in the general > context of programming languages and my personal experience of their history, > Lisp 1.5 was the most profound paradigm shift. Yes. But I was comparing to my personal influence. :) > If someone were to come to me today and ask me, "Should I learn Scheme or Common > Lisp?" I'd have a tough choice, but I think Scheme would win -- it's a much > better learning tool than Common Lisp. But two of Scheme's three major > innovations -- lexical scoping and continuations -- made it into Common Lisp. Common Lisp doesn't have continuations. >> APL lives on as J, K and Q. I still wait for the day I get to see the >> K source code... it must be an amazing piece of code. > The open source version of "APL" is A+ or aplus, depending on who you talk to. > I'm on their mailing list -- it was very quiet for most of the year but it > seems to be picking up in activity. It's in Debian. There was a lot of development between A+ and Q, though. Alone the fact that A+ doesn't use ASCII is a big turn off. I think Q is far more elegant than A+, but I didn't play a lot with it. >> Further languages that changed my way to think about programming: >> - Prolog (mainly the different evaluation mechanism) > Apparently there has been a rebirth of activity in Prolog and Prolog-like > languages such as Mercury and the mini-Kanren package built in Scheme. I > thought it was dead. I made a feeble attempt to learn it, but it made my head > hurt. It was way too far away from my "roots" as a numerical practitioner. Once you got the evaluation model of Prolog, it's just Lisp in a different syntax. :-) >> - Haskell (typing done right?) > There are just enough useful pieces of software written in Haskell (and OCaml, > too, for that matter) that you need to "have it on your hard drive". If those > pieces of software get useful enough, they'll need to be re-written in Java, > C++ or C#. :) Have an example of that? >> I'm sure I forgot something. > Well, the whole ML family, of which OCaml seems to be the only thriving member. Well, I covered functional programming with Haskell, and other than that, the things I consider convincing about ML are the generally very good implementations and maybe their module support. Not truly world-changing to me, though. Maybe it's because I learned about Haskell before ML... -- Christian Neukirchen http://chneukirchen.org