News for hackers. Stuff that matters.
Re:Ten Years of Git: An Interview with Linus Torvalds by vilhelm_s on Monday April 06, 2015 @05:06PM (#9330701)
But perhaps a better type system can still be helpful. E.g. F# supports units of measure, so you can declare 'dollar' as a unit, and write

    let calc_total (meal_cost : float<dollar>) (tax : float) (tip : float) =
       meal_cost * (1.0+tax) * (1.0+tip)
This way it is at least clear from the types that the numbers are meant to be multiplied, not added.
Re:Ten Years of Git: An Interview with Linus Torvalds by S4M on Monday April 06, 2015 @05:12PM (#9330731)
I never used F#. In your example, would the code:

    meal_cost + tax
trigger a compiler exception?
Re:Ten Years of Git: An Interview with Linus Torvalds by vilhelm_s on Monday April 06, 2015 @05:24PM (#9330777)
Yes, you get

    The type 'float' does not match the type 'float<dollar>'
There is an online typechecker, so one can try out F# without installing it: http://www.tryfsharp.org/Learn/scientific-computing#units-of...
Re:Ten Years of Git: An Interview with Linus Torvalds by S4M on Monday April 06, 2015 @05:28PM (#9330798)
And how about:

    meal_cost * tax
?

This one makes sense in the real world, so I suppose an advanced type system allows the programmer to specify what operations are legit or not across different types.

Re:Ten Years of Git: An Interview with Linus Torvalds by vilhelm_s on Monday April 06, 2015 @06:51PM (#9331162)
One could imagine systems for that, but units of measure doesn't do any customization of different operations. It's just unit checking for for arithmetic, like in high-school math.