The MIT Lightweight Languages Workshop

fxn points to this "article from Dr. Dobbs's Jounal on the recent Lightweight Languages Workshop, held at MIT's AI Lab." This is a nice followup to Simon Cozen's report on the conference, posted here in November.

Not enough Ruby

[pong]

Honestly, I think Perl is the ultimate light weight language, as it is not (in my opinion) suited for anything but lightweight tasks. Use perl for the things that you used to do with sed, awk and bash, and perhaps a few things that you didn't bother doing with those tools because it is too messy.

Joking aside, I really recommend developers who care about efficiency (not runtime :-)) to learn ruby. Its extremely expressive and very maintainable. Too bad it got so little attention at that conference.

Re:Not enough Ruby

[pthisis]

Python fits the efficiency bill very will, and the psyco project is taking steps to address runtime for the speed freaks out there. Pretty interesting, although it's not ready for serious use quite yet.

http://sourceforge.net/projects/psyco/

Personally I prefer Python from a programmer efficiency standpoint, but ruby's good too.

Sumner

Not enough Ruby & Curl

[metlin]

True. In fact, I'm surprised that lesser known languages like Lambda, Curl & Proto were given more importance than Ruby.

But on a different note, Curl is a pretty interesting language, and has a lot of features not seen in some of the design specific high level languages (like OO).

Not only is it a wierd mix of content and design, it has some wierder still means of accessing databases -- makes a HTTP/HTTPS request to an PHP/ASP/JSP/Perl file which performs the operation.

Although the interfaces generated are pretty nifty, these mostly look pretty only on IE.

And these guys have a business model that tracks the clients to see which servers using curl were accessed by those. Some privacy policy these ppl must have...

Re:Not enough Ruby

[Matts]

What I'd really personally like to see is ruby catch up with perl (and python) performance wise. I haven't checked the latest 1.7 betas/cvs, but I think it's still significantly slower on things like hash lookups. Hopefully a parrot backend for everyone would even out the scores in things like that, and allow for universal just-in-time compilation.

I'm really looking forward to parrot getting real-code front ends.

Perl is the odd man out

[tshoppa]

Interesting Dr Dobbs article, especially on the attitudes towards Perl and the attitudes by Perl developers.

It's strange to me that Perl was even thought of as a "lightweight language". My definition of "lightweight" is that the whole language spec fits in a book no thicker than K&R 1st edition. This would barely begin to describe Perl's oddities (all reference books are filled with footnotes like "unless $/ is defined to be "\n\n" and the -ip qualifier was used to invoke Perl").

Perl is also very different because it really is an evolved language. It's a strong testament to the overall framework that the original language could be extended into what we have today, and that the result is so widely used (even though it may not be ideal for anything, it's usable for many things).

OTOH it also shows the signs of evolution (see my comment about all the footnotes having to deal with special cases for backwards- compatiblity.)

What is a lightweight language?

[angel'o'sphere]

All languages mentioned in the article are "interpreted" languages. But I asume there is more which is needed to define a language as lightweight.

IMHO a language should be easy to learn, consistent in the forms how control structures and declarations are applied and the various concepts of the language should fir orthogonal together.

Perl, e.g. is not a easy to learn language.
Especial if you neither know sed, awk etc. (who even of the /. ers is today fluent in those anyway?)

So: can a language wich has a runtime environment of several tenth of megabytes be considered lightweight?

I would argue light weight languages do not even exist. Probably AppleScript or HyperCard is the only language I would consider leightweight.

Well, LISP and Co., why not. But those are for me hard to read and hard to work with, wrong mindset I asume.

Just my 2 cents (now we have those in Europe also!, yeah!)

Regards,
angel'o'sphere