FP in D (today)

[Walter Bright]

On 2/23/2012 6:50 AM, Gour wrote:
> I'm curious if
> anyone more experienced with using FP paradigm in D can comment what
> works nicely and is there anything important missing in today's
> implementation of D?

The main issue with FP in D today is it needs to be pervasively used in Phobos. 
A large part of Phobos should be marked as pure, but is not, making it difficult 
to use Phobos from FP code.

For example, Object.opEquals should be pure, nothrow, @safe.

The inheriting purity thing helps significantly with that.