4 ms·
I don't understand enough to know how significant this work is. They present a system to infer dependent types, which, if it works well, would solve at least pa
by MichaelGG 15y ago
I don't understand enough to know how significant this work is. They present a system to infer dependent types, which, if it works well, would solve at least part of the problem of dependent types as I understand it: having to write out tons of type annotations all over the place.
Having never used dependent types, perhaps I'm just being romantic, but it seems that even having some constraints on just a few types (numerics, arrays/lists, perhaps strings), would be a great boon, and I hope one day we'll have a "commercial" (that is, useful with great tooling) language that supports dependent types.
- leif 15y agoI'd be happy with an extension to an existing language, that added a little new syntax. So you give the extended source to the new tool, it verifies type safety, and then strips the new syntax elements and passes it off to the original language's compiler. This can't be that hard, can it?
- chrisaycock 15y agoMicrosoft Research created Spec#, which extends C# to have Eiffel-style contracts (pre- and post-conditions) specified above a function body. While the requires and ensures aren't dependent types, they do allow for extended static checking via Microsoft's Z3 theorem prover. http://en.wikipedia.org/wiki/Spec_Sharp http://en.wikipedia.org/wiki/Spec_Sharp
- Rickasaurus 15y agoAlso check out F* http://research.microsoft.com/en-us/projects/fstar/ http://research.microsoft.com/en-us/projects/fstar/