Interview with Samantha Kleinberg on CL-GODB, Common Lisp & Bioinformatics [LWN.net]

Interesting languages

Posted Feb 6, 2006 19:39 UTC (Mon) by ncm (subscriber, #165) [Link]

It seems to me that if you're going to put the work into learning an interestingly different language, you might as well learn one that benefits from a few of the lessons learned in (say) the last three or four decades. OCaml, Haskell, and even Standard C++ qualify. The last has the advantage of being industrially useful, besides.

Whichever you choose, expect to learn enough to be able to write a useful program in a few days or weeks, but to spend years getting really proficient. Any language that doesn't take years to master hasn't enough to teach you to be worth picking up in the first place.

Interesting languages

Posted Feb 6, 2006 22:09 UTC (Mon) by fergal (subscriber, #602) [Link]

I'm not sure how C++ qualifies as "interestingly different" unless you're going to go off into heavy-duty compile-time programming with advanced templates.

Also, do you think that Lisp hasn't benefited from the lessons learned in the last 40 decades? Lisp's syntax hasn't changed (much) because it is extensible. When someone thinks of a new way of doing things they don't need to rewrite the compiler and break all the old code to do it. That doesn't mean that everyone is stuck programming Lisp like it's still 1958.

At the rate we are going, Lisp will still be interesting in 100 years ...

Posted Feb 7, 2006 0:25 UTC (Tue) by brugolsky (subscriber, #28) [Link]

Better type systems, lazy evaluation, calculi of every variety -- all of these things are great, and can be simulated in Lisp with varying degrees of effort. Meanwhile, every newfangled language seems to end up with some meta-programming front-end that looks half-baked compared the Lisp family. Been there before with Prolog. Prolog has some interesting features; those features are readily implemented in a Lispi library.

Interesting languages

Posted Feb 14, 2006 10:29 UTC (Tue) by jschrod (subscriber, #1646) [Link]

Any language that doesn't take years to master hasn't enough to teach you to be worth picking up in the first place.

Interestingly. I would put it exactly the other way round: Any language that takes years to master is not only not worth picking up in the first place, it is actually dangerous and should be shunned to the bottom of a dark pit. Needing years to master means that the semantics are too difficult, and thus are error-prone, and this means that it's too easy to make security-related errors.

That said, neither OCaml nor C++ fit the `years to learn' qualifier. (Though I'm tempted to claim that C++ is a dangerous two-sided sword and beginners make things with it that makes me sweat. I would also argue that its semantics are too complex for its own good.) I don't know Haskell from big projects, but it doesn't look so hard to learn either, e.g., no big libraries with common idioms that I've seen. We should not forget that the typical idioms and big or complex libraries take up the most part of the time in learning a new language.

And Common Lisp is definitively in the league of one should learn it, this one language teaches almost all lessons of the last three or four decades.

Cheers, Joachim