20 Nov 2012 04:55

## code length in Haskell, a comparison

```There is some interesting data in the article at:

Code Length Measured in 14 Languages
http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/

basically comparing program lengths in various languages, and some ensuing discussion of how this
relates to language expressiveness, etc.
(He does all of his analysis in Mathematica, which is the goal of the article.)

It is interesting to see how well Haskell showed in the data; and it would also be interesting to see how well
it could replicate the analysis example which was a nice example of web data scraping!

The data is the length of a series of programs written in a number of languages (data from: http://rosettacode.org/wiki/Rosetta_Code).
(The columns don't map well to text only, Haskell column marked with ≡≡≡≡ (Why doesn't this list
support HTML?)).
See nicer version at: http://pastehtml.com/view/ciy7woohv.rtxt

The average for Haskell of 1.89 means that on the average the same program in Haskell takes ~2x in the other languages.
Given the correlation of size to clarity, complexity, effort, and errors, this is a good thing!

Code Size relative to Mathematica
Larger numbers indicate that the language on top needs more code.

C	C++	Fortran	Java	CLisp	Python	C#	JavaScript	R	MATLAB	Clojure	Pascal	Haskell	Ruby	Average
???????????	????	????	????	????	????	????	????	????	????	????	????	????	≡≡≡≡	????
Mathematica	17.0	9.1	8.1	6.4	6.3	7.2	6.4	5.0	3.2	3.2	1.6	5.8	3.5	5.2	6.29
Ruby    	2.7	1.8	1.9	1.3	1.1	1.1	1.5	1.0	0.7	0.9	0.4	1.4	0.7		1.27
Haskell	3.6	2.7	2.5	2.0	1.6	1.7	2.2	1.5	1.1	1.5	0.7	2.1		1.4	1.89
Pascal   	2.2	1.5	1.2	0.8	0.8	0.8	1.0	0.8	0.5	0.6	0.2		0.5	0.7	0.89
Clojure	8.8	5.3	5.2	3.6	3.7	3.3	3.8	2.5	1.9	2.9		5.0	1.5	2.6	3.85
```

20 Nov 2012 05:34

### Re: code length in Haskell, a comparison

```I am leery of code comparisons (but not Timothy Leary of them).

Clojure being a JVM language has the advantage of the massive Java
class libraries.

If Haskell could tie in fairly seamlessly to the Java class libraries ...

If more developers learned the need for finer grained abstraction ...

On Mon, Nov 19, 2012 at 7:55 PM, Gregory Guthrie <guthrie <at> mum.edu> wrote:
> There is some interesting data in the article at:
>
>    Code Length Measured in 14 Languages
>     http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/
>
> basically comparing program lengths in various languages, and some ensuing discussion of how this
relates to language expressiveness, etc.
> (He does all of his analysis in Mathematica, which is the goal of the article.)
>
> It is interesting to see how well Haskell showed in the data; and it would also be interesting to see how well
it could replicate the analysis example which was a nice example of web data scraping!
>
> The data is the length of a series of programs written in a number of languages (data from: http://rosettacode.org/wiki/Rosetta_Code).
> (The columns don't map well to text only, Haskell column marked with ≡≡≡≡ (Why doesn't this list
support HTML?)).
>   See nicer version at: http://pastehtml.com/view/ciy7woohv.rtxt
>
> The average for Haskell of 1.89 means that on the average the same program in Haskell takes ~2x in the other languages.
> Given the correlation of size to clarity, complexity, effort, and errors, this is a good thing!
>
```

20 Nov 2012 05:38

### Re: code length in Haskell, a comparison

I find myself wondering where Rebol would stand in this.

On Nov 19, 2012 8:35 PM, "KC" <kc1956 <at> gmail.com> wrote:
I am leery of code comparisons (but not Timothy Leary of them).

Clojure being a JVM language has the advantage of the massive Java
class libraries.

If Haskell could tie in fairly seamlessly to the Java class libraries ...

If more developers learned the need for finer grained abstraction ...

On Mon, Nov 19, 2012 at 7:55 PM, Gregory Guthrie <guthrie <at> mum.edu> wrote:
> There is some interesting data in the article at:
>
>    Code Length Measured in 14 Languages
>     http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/
>
> basically comparing program lengths in various languages, and some ensuing discussion of how this relates to language expressiveness, etc.
> (He does all of his analysis in Mathematica, which is the goal of the article.)
>
> It is interesting to see how well Haskell showed in the data; and it would also be interesting to see how well it could replicate the analysis example which was a nice example of web data scraping!
>
> The data is the length of a series of programs written in a number of languages (data from: http://rosettacode.org/wiki/Rosetta_Code).
> (The columns don't map well to text only, Haskell column marked with ≡≡≡≡ (Why doesn't this list support HTML?)).
>   See nicer version at: http://pastehtml.com/view/ciy7woohv.rtxt
>
> The average for Haskell of 1.89 means that on the average the same program in Haskell takes ~2x in the other languages.
> Given the correlation of size to clarity, complexity, effort, and errors, this is a good thing!
>

Massive Snip!

>
> Overall Ranking:
>    Clojure              0.41
>    R                    0.85
>    MATLAB               1.14
>    JavaScript           1.27
>    Ruby                         1.36
>    Common Lisp          1.56
>    Python               1.61
>    C++                  1.68
>    C#                   1.85
>    Pascal               1.86
>    Fortran              2.33
>    C++                  2.58
>    C                    4.09
> -------------------------------------------
>
> _______________________________________________

--
--
Regards,
KC

_______________________________________________
```_______________________________________________
```
20 Nov 2012 06:10

### Re: code length in Haskell, a comparison

```Darren Grant <therealkludgy <at> gmail.com> wrote:

> I find myself wondering where Rebol would stand in this.

Or APL.

Greets,
Ertugrul

--

--
Not to be or to be and (not to be or to be and (not to be or to be and
(not to be or to be and ... that is the list monad.
```
```_______________________________________________
```
20 Nov 2012 14:16

### Re: code length in Haskell, a comparison

They also have other comparisons at the referenced site, including for different sizes of programs, and for counting characters or tokens instead of lines over each of these program example groups.

The data source does include APL & REBOL (& 483 different languages!), so one could run this analysis with most any languages you liked.

Analysis from: http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/

Data      from: http://rosettacode.org/wiki/Rosetta_Code

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

> I find myself wondering where Rebol would stand in this.

Or APL.

Greets,

Ertugrul

-------------------------------------------

```_______________________________________________
```
20 Nov 2012 14:32

### Re: code length in Haskell, a comparison

Why C++ appears twice in the Overall rank?

2012/11/20 Gregory Guthrie
There is some interesting data in the article at:

Code Length Measured in 14 Languages
http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/

basically comparing program lengths in various languages, and some ensuing discussion of how this relates to language expressiveness, etc.
(He does all of his analysis in Mathematica, which is the goal of the article.)

It is interesting to see how well Haskell showed in the data; and it would also be interesting to see how well it could replicate the analysis example which was a nice example of web data scraping!

The data is the length of a series of programs written in a number of languages (data from: http://rosettacode.org/wiki/Rosetta_Code).
(The columns don't map well to text only, Haskell column marked with ≡≡≡≡ (Why doesn't this list support HTML?)).
See nicer version at: http://pastehtml.com/view/ciy7woohv.rtxt

The average for Haskell of 1.89 means that on the average the same program in Haskell takes ~2x in the other languages.
Given the correlation of size to clarity, complexity, effort, and errors, this is a good thing!

Code Size relative to Mathematica
Larger numbers indicate that the language on top needs more code.

C       C++     Fortran Java    CLisp   Python  C#      JavaScript      R       MATLAB  Clojure Pascal  Haskell Ruby    Average
???????????        ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ≡≡≡≡    ????
Mathematica     17.0    9.1     8.1     6.4     6.3     7.2     6.4     5.0     3.2     3.2     1.6     5.8     3.5     5.2     6.29
Ruby            2.7     1.8     1.9     1.3     1.1     1.1     1.5     1.0     0.7     0.9     0.4     1.4     0.7             1.27
Haskell 3.6     2.7     2.5     2.0     1.6     1.7     2.2     1.5     1.1     1.5     0.7     2.1             1.4     1.89
Pascal          2.2     1.5     1.2     0.8     0.8     0.8     1.0     0.8     0.5     0.6     0.2             0.5     0.7     0.89
Clojure 8.8     5.3     5.2     3.6     3.7     3.3     3.8     2.5     1.9     2.9             5.0     1.5     2.6     3.85
MATLAB  3.6     2.4     1.8     1.1     1.4     1.1     1.7     0.9     0.8             0.3     1.6     0.7     1.1     1.42
R               4.7     3.3     2.4     1.9     1.8     1.7     2.1     1.5             1.3     0.5     2.2     0.9     1.4     1.98
JavaScript      2.8     2.1     1.9     1.2     1.2     1.1     1.6             0.7     1.1     0.4     1.3     0.7     1.0     1.31
C#              2.0     1.4     1.3     0.9     0.8     0.8             0.6     0.5     0.6     0.3     1.0     0.5     0.7     0.87
Python          2.2     1.6     1.5     1.1     0.9             1.2     0.9     0.6     0.9     0.3     1.3     0.6     0.9     1.07
Common Lisp     2.8     1.8     1.6     1.3             1.1     1.3     0.8     0.6     0.7     0.3     1.3     0.6     0.9     1.16
Java            2.1     1.4     1.5             0.8     0.9     1.1     0.8     0.5     0.9     0.3     1.2     0.5     0.8     0.98
Fortran 1.4     1.0             0.7     0.6     0.7     0.8     0.5     0.4     0.6     0.2     0.8     0.4     0.5     0.66
C++             1.4             1.0     0.7     0.6     0.6     0.7     0.5     0.3     0.4     0.2     0.7     0.4     0.6     0.61
C                       0.7     0.7     0.5     0.4     0.5     0.5     0.4     0.2     0.3     0.1     0.5     0.3     0.4     0.41
????????????  ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ????    ≡≡≡≡    ????
Overall:        4.1     2.58    2.3     1.68    1.56    1.61    1.85    1.27    0.85    1.14    0.41    1.86    0.83    1.29

Overall Ranking:
Clojure              0.41
R                    0.85
MATLAB               1.14
JavaScript           1.27
Ruby                         1.36
Common Lisp          1.56
Python               1.61
C++                  1.68
C#                   1.85
Pascal               1.86
Fortran              2.33
C++                  2.58
C                    4.09
-------------------------------------------

_______________________________________________

--

Feng, Boqun
School of Software，Shanghai Jiao Tong University

```_______________________________________________
```
20 Nov 2012 23:36

### Re: code length in Haskell, a comparison

```
On 20/11/2012, at 4:55 PM, Gregory Guthrie wrote:

> There is some interesting data in the article at:
>
>   Code Length Measured in 14 Languages
>    http://blog.wolfram.com/2012/11/14/code-length-measured-in-14-languages/
>
> basically comparing program lengths in various languages, and some ensuing discussion of how this
relates to language expressiveness, etc.
> (He does all of his analysis in Mathematica, which is the goal of the article.)

I'm not sure how interesting it actually is.
Let's study just one example: the digit sum problem.

"This task is to take a Natural Number in a given Base and return the sum of its digits."

We are not given any bounds on the size of the Natural Number
nor any bounds on the Range.  Arguably, the code for almost all of the
programming languages shown is *wrong* due to making unwarranted

digsum base = f 0 where
f a 0 = a
f a n = f (a+r) q where
(q,r) = n `divMod` base

main = print \$ digsum 16 255 -- "FF": 15 + 15 = 30

Since some other examples assume the base is 2..16, so may we.
In that case,

import Data.Char
import Numeric
digsum b n = sum . map digitToInt \$ showIntAtBase b intToDigit n ""

will do the job, and as an exercise in golfing myself,

import Numeric
digsum b n = sum . map fromEnum \$ showIntAtBase b toEnum n ""

will also do it, and should work for larger bases.

In this particular test, Mathematica happens to score well
because it already has IntegerDigits[number{, base}?] in its
library.

So at least this exercise is much more about "what happens to
be available already in the library for this language" than
about anything intrinsic to the language.

There is worse.  Some versions read test data, while some have
a small number of test cases built in, and they do not agree
about which test cases are provided.

Looking at some other problems, some tasks have versions that
use code that is published on the web but not part of the
language standard and not included (or counted!) at the Rosetta
site itself.

There are lots of things you can learn by looking at the
Rosetta site.  I am not sure that the compactness of languages
is one of them.
```

Gmane