More Readable Use Site extension Methods for java
reinier at zwitserloot.com
Sun Dec 12 21:33:25 PST 2010
This idea isn't exactly new, it's been mentioned countless times, here and
An implementation would surely help in analysing cost/benefit for this
feature, but, as Brian said, there's also the simple concern of new feature
complexity. Most likely a beautiful patch with a thousand tests associated
with it still won't make it into JDK8.
See for example the elvis operator, where IIRC a patch was ready to go, but
ended up nixed off the coin list because the feature did not pull its weight
compared to the complexity it adds (personally I happen to disagree with
that decision but didn't feel strongly enough about it, at least at the
time, to experiment). In other words, many features will not be implemented
because any feature makes java (the language) more complex, it's not always
due to lack of manpower to implement and test the feature. In fact, I'd go
so far as to say it virtually never is due to lack of manpower; instead,
there's always a lack of manpower to more thoroughly figure out which
features are worth their cost and which aren't.
If you're convinced that Brian is wrong and that this feature _IS_ worth its
complexity cost, a first step towards showing this is to build a prototype
and clean up some 'real life' code (pick some well known FOSS project, like
a bunch of hibernate code, or jetty, or glassfish, or whatever strikes your
fancy), and show both the feature in action and the diffs that represent
significant savings in readability.
On Tue, Dec 7, 2010 at 8:26 PM, Llewellyn Falco <isidore at setgame.com> wrote:
> ok, let me open this offer up to the list then.
> On Tue, Dec 7, 2010 at 11:13 AM, Brian Goetz <brian.goetz at oracle.com>
> > Thanks for the offer! I personally am more involved in the specification
> > than the implementation, but there are those involved in implementation
> > might well want to take you up here. And lambda-dev *is* the right place
> > make that offer!
> > Cheers,
> > -Brian
> > On 12/7/2010 1:38 PM, Llewellyn Falco wrote:
> >> ok, maybe coding would be a good place to start. would it be possible to
> >> do
> >> some pairing?
> >> i've remote paired a lot, it's pretty easy to do. We can do either
> >> advanced
> >> pairing via a ec2 image, or simple pairing via skype screen share.
> >> I would suggest we start with a time boxed 1 hour.
> >> We can either do whatever you are currently working on, or we could do
> >> write
> >> some unit tests, which should be rather easy as lambdas fall under the
> >> category of "theory based testing"
> >> Does this sound good?
> >> Llewellyn Falco
> >> skype: isidore_us
More information about the lambda-dev