Article about the expressiveness of languages with D included as one of the contestants.

http://redmonk.com/dberkholz/2013/03/25/programming-languages-ranked-by-expressiveness/

I tend to agree with the first comment to the article though :)

/Jonas

On Tuesday, 2 April 2013 at 07:59:17 UTC, Jonas Drewsen wrote:
> Article about the expressiveness of languages with D included as one of the contestants.
>
> http://redmonk.com/dberkholz/2013/03/25/programming-languages-ranked-by-expressiveness/
>
> I tend to agree with the first comment to the article though :)
>
> /Jonas

And me with the one about Go.

On 4/2/2013 12:59 AM, Jonas Drewsen wrote:
> Article about the expressiveness of languages with D included as one of the
> contestants.
>
> http://redmonk.com/dberkholz/2013/03/25/programming-languages-ranked-by-expressiveness/

It's an interesting metric, but there are too many obvious confounding variables to assume that expressiveness has the first order effect.

On 04/02/2013 09:59 AM, Jonas Drewsen wrote:
> Article about the expressiveness of languages with D included as one of the contestants.

Personal feeling here -- there's a difference between how expressive a language can be (even, how expressive it can _easily_ be) versus how expressively programmers tend to use it.

I think my own use of D tends to be heavily biased by my background in C/C++ and my lack of training in more expressively-focused development styles.  D allows me to write in those paradigms I feel comfortable with -- and so my use of it is almost certainly less expressive than it could be.

That feeling is supported by how wide D's error bars are in those plots -- that diversity may well reflect the number of styles of programming one can adopt within the language.  I'm surprised that the extreme lower values for the statistic still seem high relative to other languages, but that in turn might reflect the state of development of the language, with new features being added fairly regularly to the standard library (probably larger commits).

I also have a strong feeling that LOC per commit reflects too many different factors to be really reliable as a comparison, e.g. it probably depends quite strongly on the age/maturity of a project, the rate of development, and other factors.

Reading some later posts on the same blog, the author acknowledges some of these
kinds of complications:
http://redmonk.com/dberkholz/2013/03/26/what-does-expressiveness-via-loc-per-commit-measure-in-practice/

On 04/02/2013 11:15 AM, Walter Bright wrote:
> It's an interesting metric, but there are too many obvious confounding variables to assume that expressiveness has the first order effect.

Between your response and mine, I think we have a rather good illustration of this for the English language, never mind programming ... :-)

Jonas Drewsen:
> Article about the expressiveness of languages with D included as one of the contestants.
>
> http://redmonk.com/dberkholz/2013/03/25/programming-languages-ranked-by-expressiveness/

I think D is quite expressive:
http://forum.dlang.org/thread/zdhfpftodxnvbpwvklcv@forum.dlang.org

Bye,
bearophile

On Tuesday, 2 April 2013 at 07:59:17 UTC, Jonas Drewsen wrote:
> Article about the expressiveness of languages with D included as one of the contestants.
>
> http://redmonk.com/dberkholz/2013/03/25/programming-languages-ranked-by-expressiveness/
>
> I tend to agree with the first comment to the article though :)
>
> /Jonas

Yep, the sorting seems quite random to me, AFAIK Vala is nothing special yet it is ranked very high in this article..

renoX

On 04/02/2013 03:24 PM, renoX wrote:
> Yep, the sorting seems quite random to me, AFAIK Vala is nothing special yet it is ranked very high in this article..

To be fair, the author does say that results for what he calls "third tier"
languages (like Vala) should be considered with a great deal of skepticism:
http://redmonk.com/dberkholz/2013/03/26/what-does-expressiveness-via-loc-per-commit-measure-in-practice/

On 4/2/2013 2:53 AM, Joseph Rushton Wakeling wrote:
> I also have a strong feeling that LOC per commit reflects too many different
> factors to be really reliable as a comparison, e.g. it probably depends quite
> strongly on the age/maturity of a project, the rate of development, and other
> factors.

Consider also that this LOC numbers are not lines of code - they're also lines of comments! D's ddoc encourages writing considerably more lines of comments than C does.

On Tuesday, 2 April 2013 at 17:33:13 UTC, Walter Bright wrote:
> On 4/2/2013 2:53 AM, Joseph Rushton Wakeling wrote:
>> I also have a strong feeling that LOC per commit reflects too many different
>> factors to be really reliable as a comparison, e.g. it probably depends quite
>> strongly on the age/maturity of a project, the rate of development, and other
>> factors.
>
> Consider also that this LOC numbers are not lines of code - they're also lines of comments! D's ddoc encourages writing considerably more lines of comments than C does.

Not to mention that idiomatic D bracing style adds more lines.