Optional -> OptionalResult

Zhong Yu zhong.j.yu at gmail.com
Tue Jun 4 14:26:06 PDT 2013


"Option" sounds like the producer has a choice, it can choose to
produce something or nothing based on some private/undisclosed
reasons. That connotation may mislead people into using it for the
wrong reasons, e.g. as input type.

On Tue, Jun 4, 2013 at 3:36 PM, Clayton Wohl <claytonwohl at gmail.com> wrote:
> Why would "Map<Optional<List<Optional<String>>>" be more of a concern than
> any other ridiculous combination of generics nesting? How about
> List<Vector<Stack<AtomicReference<ThreadLocal<Integer>>>>>?
>
> Usually, devs don't nest optionals inside of lists/maps. I'm sure it will
> happen in some edge case, but in the Scala code or the Functional Java code
> bases I've worked in, people do not do that, just like they don't do lots
> of other crazy combinations of generics that are technically possible in
> Java 5+.
>
> For the name, I like Option/Optional, however, that's a relatively minor
> detail and I will be happy with whatever the JDK team chooses.
>
> Stephen, the reasoning for an Optional type with map/flatMap/forEach
> functionality is very convincing and I don't believe it's appeal is limited
> to elite FP experts, which I am definitely not. I've seen lots of junior
> Java devs appreciate the same functionality in Functional Java. Your
> opposition to this does not seem to have a grounding in reason.
>
> BTW, the new date/time library you've worked on looks amazing. I've been
> using the plug-in version in my personal projects for years, and I am
> excited to see it broadly adopted into the core JDK 8.
>
>
> On Tue, Jun 4, 2013 at 10:13 AM, Stephen Colebourne
> <scolebourne at joda.org> wrote:
>> Optional is, and will continue to be, a key debating point for those
>> arguing for more FP. The pressure to make it more powerful will
>> continue, especially given similar named features in other languages.
>>
>> Meanwhile I continue to fear Map<Optional<List<Optional<String>>> and
>> the like. The generics type system in Java simply isn't good enough to
>> be doing anything like that, even if it were remotely desirable.
>


More information about the lambda-dev mailing list