What's the most perfectly succinct programming language?

[Bert Bates]

Some people contend that certain languages, like, er, ahem... Java, might be too verbose.

Some people contend that other languages, like Perl, are probably too terse.

What's YOUR favorite language from the perspective of succinctness vs. verbosity?

[Hussein Baghdadi]

Clojure then Haskell.

[Ernest Friedman-Hill]

English.

[Jeanne Boyarsky]

Ernest Friedman-Hill wrote:English.

I don't know. English (requirements) tend to be quite ambiguous.

[Hussein Baghdadi]

Kanji is considered succinct or verbose?

[Stephan van Hulst]

AFAIK Kanji is a script.

[Bert Bates]

..bunch of wise-acres...

How about 'most perfectly succinct *programming* language'

[Jeanne Boyarsky]

Bert Bates wrote:..bunch of wise-acres...

That's what you get for asking serious question in MD

Bert Bates wrote:How about 'most perfectly succinct *programming* language'

Is picking a DSL cheating? Domain specific languages are designed against the domain so a lot can be abstracted.

[Pat Farrell]

Iverson's APL was very succinct. Perl is much more dense that Java, but then again, nearly everything is.

Many folks will argue that LISP is the most expressive language.

[Bert Bates]

Recently I've had the opportunity to talk with some 20-something, alpha-geek programmers. One recurring theme I hear from them is that they like languages that are less verbose than Java. These languages also *tend* to be dynamic languages (which they sometimes refer to as 'scripting' languages, and I think those definitions aren't precise and are shifting as we speak.) Some examples are CoffeeScript and Ruby.

Some other recurring themes are:

- Perl is TOO succinct
- They're looking for a quality - that's still a little elusive to me - that they call 'expressiveness'.

There are a lot of factors here, but one that we've talked about, and that I'm bringing up here, is that perhaps human brains are naturally 'tuned' to an optimum 'sweet spot' of succinctness. Hence this thread...

One, possibly related, concept that has been endlessly debated here at the ranch is the "proper" formatting of curly braces. At some level, the various styles create differing degrees of succinctness.

[Pat Farrell]

Bert Bates wrote:They're looking for a quality - that's still a little elusive to me - that they call 'expressiveness'.

I think that expressiveness is when you can apply an operator or method and know what it will do by looking at it. And when you combine chains of operators/methods, they work in intuitive ways.

The functional languages have it, but they look too foreign for most classically trained programmers.

I first saw it not in a language, but in an operating system: Tops-20/Tenex. If you could take a thing A and apply a function to it, say F(A) and get a result, and there is another thing B and another function G that has G(B) work, you can take G(F(A)) and have it work the way you expect, without looking up any manuals. It was liberating.

[Hussein Baghdadi]

Indeed.
After coding in Clojure and Haskell, I feel the functional programming language are really succinct. Composing functions is the first thing comes to my mind.

[Paul Clapham]

I used to program in APL, which was pretty darn succinct. That was 30 years ago; recently I saw an example of APL which was much more succinct than I remember APL being.

[Pat Farrell]

Paul Clapham wrote:That was 30 years ago; recently I saw an example of APL which was much more succinct than I remember APL being.

You sure it wasn't 40 years ago? I thought it had died out by 30 years ago.

APL programmers used to compete on how short they could make their programs. At least they did in the days when you used an IBM Selectric terminal with overtyping to make up the operators. I remember one was a box with a classic "divide" sign from arithemitic (÷) inside the box.

Those guys code do a ton of work in one line of APL.

The downside was that it was often Kleenex code: too complex to understand by anyone other than the author, and even often not by the author a few days later. So you used it once and threw it away, like a Kleenex.

[Paul Clapham]

Pat Farrell wrote:

Paul Clapham wrote:That was 30 years ago; recently I saw an example of APL which was much more succinct than I remember APL being.

You sure it wasn't 40 years ago? I thought it had died out by 30 years ago.

I did some googling and I can say that it was no earlier than 1975.

[Kaustubh G Sharma]

counting clouds or sheep

[Ryan McGuire]

Pat Farrell wrote:
APL programmers used to compete on how short they could make their programs. At least they did in the days when you used an IBM Selectric terminal with overtyping to make up the operators. I remember one was a box with a classic "divide" sign from arithemitic (÷) inside the box.

Ah yes... a divide and a box to make the 'domino'. If I recall correctly it did a least square best fit function. One sec... let me Google that. <20 seconds of Muzak/> It can do either the least squares fit or matrix inversion depending on the arguments you give it.

[Pat Farrell]

Ryan McGuire wrote:Ah yes... a divide and a box to make the 'domino'. It can do either the least squares fit or matrix inversion depending on the arguments you give it.

But how did the inversion code handle the Longley data?

p.s. nice Irish name there.

[Ted Dunning]

APL was generally state of the art with respect to matrix computation for the time. The state of the art has progressed a good bit since, then, of course.

[Juan Francisco Paulini]

I really like Python. It let you be clear in your coding, it includes lambda expressions and a great way of documenting you code...

Juan

---
Well, I posted this message, but it's clearly nothing about how "succint" python is...

Recently in a some programming quiz there was some text oriented problems to solve, one of them was getting the number of sentences that where 4 characters long, ended with 's','r','n','w',or 'k', and not containing the 'v'..

foo = 'srnwk' def prepositions(txt): preps = [w for w in txt.split() if len(w) ==4 and not 'v' in w and w[-1] in foo ] return len(preps)

It's not only succint, but succint enought to be human frienly too...

[Sean Corfield]

Pat Farrell wrote:

Paul Clapham wrote:That was 30 years ago; recently I saw an example of APL which was much more succinct than I remember APL being.

You sure it wasn't 40 years ago? I thought it had died out by 30 years ago.

APL was still very much alive and in use in the mid-80's. I encountered it during my industrial placement year at university and so fascinated by it that my final year special project was to write an APL interpreter (in Pascal). The department even bought an APL "golfball" for the Diablo teleprinter terminal in our lab so that I could do "real" APL programming with symbols :)

[Ville Myrskyneva]

Contributing to the original topic.

I don't think there is a definitive answer to the question. The correct "language" depends pretty much on what you are actually doing.

But I'll make a small comparison for PHP and Java (since I have little experience from other languages).
Making yourself a dynamic home page (or even for some company) it is much easier and faster to do it with PHP and/or JavaScript, than with Java. Also it's much easier to find a server for hosting php than java code (unless having your own server).

When you need to do lots of BL and stuff, the PHP will be like stabbing yourself in the face. (I did my share of PHP few years back, and that's how it pretty much felt when you were doing bug fixes and new features to legacy code)
Of course the legacy code was crappy as hell, but the coding wasn't any better even when doing brand new software from the scratch (teams consisted of 5-7 programmers). The maintainability was almost as horrible than before when the first iteration after the release started.

I have seen really crappy Java code also, but in Java you are kind of enforced to do certain things in smarter ways (maybe it's just the primitives). Even when Java code is done crappy, it is more maintainable than PHP.