A Bold Prediction about Programming Languages in 10 years
I have been gently scolded for my predictions about programming languages in an earlier blog post as being somewhat meek. In response I will make one prediction that is rather bold, which I left off the list due to its somewhat provocatory nature: the dominant language 10 years from now, hasn't been designed yet.
I predict with a fair amount of confidence that none of the language that exist today, will rise to become the dominant language 10 years from now. This is primarily because the people designing the current are only starting to understand how various ideas from different paradigms can be put together effectively. You can not continue to add more legs to a horse, and expect it to run faster.
I will reiterate what I think the next mainstream language will look like:
- higher-order functions
- flexible syntax (e.g. you can overload infix operators)
- static type system
- type inference
- support for generative programming (e.g. imagine if the C++ template metaprogramming system was carefully designed from the ground-up to replace the macro pre-processor)
- built in message passing concurrency
- built-in reflection API support
Here are some of the things it won't have:
- raw memory access (i.e. pointers)
- unhygenic macros - proper support for reflection
- built-in unit tests
- built-in contracts
- virtual functions
- keywords like "const" or "final" or "static".
- direct aspect-oriented programming support
It is kind of annoying to see the same pattern of language design being repeated over and over again: come up with some mediochre ALGOL based language with a bunch of features tacked on. Add a bunch more features with each release.