On Sep 15, 11:28 pm, Bob Lee <[email protected]> wrote:
>
> I don't know where multicatch stands. You'll have to ask Neal and Joe. If
> it's out, I'm sure there's a good reason.
>

That's the crux of the situation, isn't it?

Where's that reason? We keep coming back again and again to the exact
same point:

If multi-catch doesn't make it and the explanation is not publically
known, that's going to put a real crimp on community effort for future
coin endeavours. We're all aware this is a perception issue and yet
nobody is addressing this perception at all: If there's a good reason
for multicatch to get canned, then where is it? That one-liner of:
"Messes with the typesystem" is crud - the proposal specifically
mentioned why it does not in fact require messing with the type
system. Something like "We're going to revisit a lot of types in JDK8,
and we may introduce disjoint types, so we're going to wait as we may
solve it and many others in one fell swoop in JDK8" would help with
_perception_. Which is what the community you want input from is
asking for.

We're trying to help by explaining why the community is not responding
the way Joe had hoped for, and is instead moving towards things like
lombok.

In regards to the coin form: You didn't answer my question, Joe, you
just went on the defense. Why not accept proposals based on less
officious bits and more useful bits, and make decisions earlier? You
yourself said string switch took you 8+ hours to write up including
feedback. You're the coin lead, you knew it had a good chance of
making it. You're a JLS and langtools expert. You know exactly what
you want from the proposal. The people you want input from would have
needed 24 full hours or more, and would instead have the knowledge
that the effort has at best slim chances of making it in. Why are you
surprised the community isn't putting in as many proposals as you
wanted? Putting in that much effort writing excruciatingly boring JLS
spec with virtually no guarantee is of course not going to find many
takers. If you up the odds considerably by asking people to put in the
JLS efforts ONLY after giving them a tentative guarantee that if the
work is done and no surprises crop up, the proposal will far more
likely be in than be left out, *I* would definitely be a lot more
jazzed to put in that effort. I'm making a guesstimate that the rest
of the community would be similarly willing to do so.

I'll give you an example. here's strings in switch, again, this time
in 10 minutes, and this time with the -exact- same ability to make a
_preliminary_ call on its feasibility and suitability for java:

"The strings in switch proposal allows using strings in switches. The
target strings, mentioned in the case statements, must be compile-time
constants.

Such a construct is translated to existing valid java code as such:
Each target string is hashed, and the switch is replaced with an
integer switch on the input string's hashCode. Each 'case' label
contains 1 if(equals) block comparing the input string to the constant
for each string that matches that hashcode, closing with an else that
jumps to the default."


Add a small paragraph with a syntax example and a hint about how the
existing javac grammar can easily be changed to accomodate it, and
you've got all you need to make a preliminary call.

(I know that's not how you've written up string, but exact
implementation is irrelevant at the decision stage of the process, as
long as its obvious it won't be particularly difficult to find a
workable one).

Every other bit (and more) on the coin form eventually becomes
relevant, but please, pretty please, with sugar on top, start
including the decision process into the iterative process, instead of
asking everybody to iterate their way to loads of personal time before
eliminating all that hard work in one fell swoop by simply not listing
it on the shortlist. I can guarantee you more involvement from me, and
if I'm getting the pulse of this thread's participants right, I won't
be the only one.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to