Re:Experiment: Unit testing isn't enough; You need static ty...
by kingkilr on Wednesday June 20, 2012 @10:36AM
(#4137158)
I wasn't trying to catch the error of "program author is a moron who doesn't know the difference between Fibonacci and factorial". Were I trying to catch that error I would have been aware of it, and then much less likely to write the bug in the first place. This is a truism that is well accepted by testing proponents: which tests you write are incredibly important, and you need to write your tests first in order to avoid a curve fitting problem (so to speak). Any non-trivial test would have shown my function to be very broken, what type would you have used to represent that so it wouldn't compile?
Re:Experiment: Unit testing isn't enough; You need static ty...
by dons on Wednesday June 20, 2012 @10:53AM
(#4137254)
Numerical algorithms suffer from a paucity of types. So either you enrich your numerical type hierarchy, or you prove an implementation matches a model, e.g. for fibonacci http://stackoverflow.com/a/8434107/83805
Re:Experiment: Unit testing isn't enough; You need static ty...
by kingkilr on Wednesday June 20, 2012 @10:59AM
(#4137288)
I don't have a response to that, other than to say, now you know why I decided to write compilers instead of pursue a graduate degree in computer science.
Re:Experiment: Unit testing isn't enough; You need static ty...
by Peaker on Wednesday June 20, 2012 @12:35PM
(#4137738)
If you write compilers, it would be really beneficial if you had a curiosity about the state of the art in programming languages.
Re:Experiment: Unit testing isn't enough; You need static ty...
by chrisdone on Wednesday June 20, 2012 @02:01PM
(#4138217)
… And lo, PHP-2 was born into the world.
Re:Experiment: Unit testing isn't enough; You need static ty...
by Symmetry on Wednesday June 20, 2012 @01:41PM
(#4138103)
The point about a static type system not catching all errors is well taken[1], but if a programmer has mixed up factorial and Fibonacci I would expect that their tests would reflect this as well as their code. This is more the sort of thing you should rely on code reviews to spot.
[1] Isn't the old joke in Haskell "If your program compiles it probably does something someone would find useful, but not necessarily the useful thing you want"
Just as you should write good tests, that actually test for useful properties, so you should write good types -- and get useful proofs back from the compiler as a result.