smalltt

The hype of Blockchain is an astounding thing, and the cascading hype for systems merely because they are based off people in the Blockchain space is equally as befuddling. Try https://github.com/AndrasKovacs/smalltt if you want a system that considers the things modern systems care about: elaboration and unification.

I'm curious if you have any insight into what a realistic lower limit to Idris compile time speed looks like. Right now even moderately sized Idris programs can be quite slow to compile, slower than even other dependently-typed languages (see e.g. https://github.com/AndrasKovacs/smalltt). How much of this is intrinsic to Idris' language design and how much of it is due to lack of optimization in the implementation?

Not an expert, but I've played with these languages. The short answer is yes this is possible, but it is also possible already in non-dependently typed languages to do so. Moreover in practice it's not very common to actually evaluate your functions on large literal inputs at compile time. Indeed it's not very common to evaluate the function on a single value as such at all. Instead it is better to think of your functions as abstract rewrite rules, that allow you to substitute one side with the other. Unlike at runtime, recursive functions do not always need to recurse until their base case. Therefore it's not actually very common to e.g. have a recursive function recurse many times at compile time. However, when it does happen, yeah your compile times can shoot up, but it's usually fairly obvious what's causing it. Even then, in absolute