Primitive streams and optional

[Paul Sandoz]

On Nov 27, 2012, at 4:27 PM, Doug Lea <dl at cs.oswego.edu> wrote:

> On 11/27/12 09:33, Tim Peierls wrote:
>> On Nov 26, 2012, at 10:12 PM, Sam Pullara <spullara at gmail.com> wrote:
>> 
>> I still prefer having one method that returns a "thing" that can be used,
>> uniformly and in a fluent-like manner, to do determine whether a result is
>> available or absent.
>> 
> 
> In which case (as Howard Lovatt almost said on lambda-dev list), you
> are either a closet smalltalk programmer, so want ifTrue:/ifFalse blocks,
> or a closet lambda-calculus programmer, in which case you want multiple
> continuations. Which are almost the same thing!
> So maybe you want:
> 
>  <U> U withFirst(Predicate<T> p,
>                  Function<T,U> ifPresent, Supplier<U> ifAbsent);
> 
> in addition to the plain reduction-based version.
> 

Does being a Clojure fan count ? :-)

I referred to "thing" as i think we are getting bogged down a little in this discussion.

If we had Stream.Result (or java.util.stream.Result) instead that kind of walked and quacked somewhat like the current j.u.Optional would that be less contentious?

Paul.


>> Me, too, if by "thing" you mean "Optional". It solves so many problems:
> 
> Tim: I love Optional with Value types. Vastly better than boxing.
> I hate it as a half-object/half-value compromise.
> If/when Java supports true value types, I will be first in
> the "let's do Optional right" line.
> But until then, I'm still dead set against it.
> 
> -Doug
> 
> 
>