[jamwt]

At bu.mp, for our main backend repository `sloccount` says:

    python:       22961 (67.22%)
    ansic:         6338 (18.55%)
    haskell:       4378 (12.82%)
    sh:             482 (1.41%)

Edit: I should note, however, that the trend over the last year has been increasing that haskell share; you get just a lot done in not a lot of code... reuse is very good.

[enigmo]

I didn't really believe Github's inflated line counts (125klocs) and was curious was sloccount would say...

I'm a bit surprised with the results. The sheer quantity of C/C++, brainless coding over a few days, was impressive. Experience buys something I guess :-/

    haskell:      65025 (88.22%)
    cpp:           6944 (9.42%)
    ansic:         1133 (1.54%)
    python:         391 (0.53%)
    java:           204 (0.28%)
    sh:               7 (0.01%)