Re: Generators/coroutines in future Ada?

[Maciej Sobczak <see.my.homepage@gmail.com>]

> Taking the example from Wikipedia:
> 
> function startAt(x)
>    function incrementBy(y)
>        return x + y
>    return incrementBy
> 
> variable closure1 = startAt(1)
> variable closure2 = startAt(5)

Note that this example does not exploit (pun intended) the fact that such variables can be created dynamically, where their exact number is not known until the program is run. Imagine an array (of length determined at run-time) of such generators. This shows that static approach of Ada is not always convenient and is not applicable to every problem.

Dmitry is right that generators can be replaced with functions that operate on explicit state, but generators are useful, because they have the same interface as stateless functions[*] and can be used as arguments in higher-level generics that accept "callable" entities without caring about whether they are stateful or not[**]. Functional languages (or at least those that claim to have some features from functional programming) are convenient for gluing stuff together this way, with obvious compromises in other areas.

[*] That is, there is some commonality between stateless and stateful functions and it might be useful to capture this commonality in some programming concept that is not present in Ada.

[**] C++ relies on such constructs in the STL library (and many others), but calls them "functors" instead. That is, a generator can be implemented as a stateful functor (and they can be created dynamically). Whether they are used wisely is another story.

-- 
Maciej Sobczak * http://www.inspirel.com