Exception transparency - lone throws (no checked exceptions)

[Stephen Colebourne]

On 13 June 2010 01:52, Jesse Kuhnert <jkuhnert at gmail.com> wrote:
> I think the fundamental question here is whether or not this is within
> the scope of this project?
Brian indicated it was a useful exploration of the design space.

> Nothing irks me more than being told I must catch this or that
> exception which I'm sure many if not all people might agree, but it
> exists already. We could come out of this with an entirely new
> language if we followed the path along to the end right?
The proposal does not remove checked exceptions. It simply recognizes
current reality (other JVM languages, lombok sneaky throws etc) and
runs with it. It you want a closure that declares checked exceptions,
simply define a SAM for it.

> You are making lots of "quotes" but I'm unclear as to whether you are
> personally responsible for the success of the project other than some
> vague references to "speaking for the common man, the general
> developer".
Language design is hard. And I'm not a language designer (never
claimed I was). My role here is to assist language designers in
finding the sweet spot where the concepts that they wish to express
are usable to a broader audience. (By definition, everyone on this
list is not an average developer, so it is always a personal judgement
call as to what is and isn't usable by that wider audience). In the
end, Oracle decides.

I should note that given project lambda is heavily based in semantics
and syntax on the FCM proposal, itself a deliberate attempt to find
the sweet spot between BGGA and CICE, I don't think it unreasonable
for me to have views here, nor argue constructively for them.

BTW, if an Oracle member of staff wants to categorically state that
this proposal is unacceptable, then please do so unequivocally. Doing
so, would terminate the discussion. Of course, it wouldn't terminate
the discussion of how the heck to express exception abstraction as
usable syntax :-)

Stephen