On Apr 27, 2010, at 11:28 AM, Kevin Wright wrote:
The latest draft document on the design of scala.swing isn't bad: http://www.scala-lang.org/sid/8
, through it doesn't show much beyond the use of closures/function
pointers.
Yes - I'm a fan of scala, except that I wish it were more coherent and
consistent in syntax. That's a personal nit though.
Even more interesting is this document on "Deprecating the Observer
Pattern", which really does bring the full power of FP to the
problem, including delimited continuations: http://infoscience.epfl.ch/record/148043/files/DeprecatingObserversTR2010.pdf
[warning: pdf]
I've come to the conclusion that FP is a degenerate case of the
Observer pattern - which is why Clojure works so nicely in an object-
oriented bytecode virtual machine.
And there's JavaFX of course! All of these are very definitely
using functional elements in their design, although the need to
interact with Swing is understandably going to restrict the
possibilities here (less so in the deprecating observers document)
Yep. All of this is awesome. And starts to make up for some of
Java's flaws. We're working on a side project called "noop" which is
a different take - trying not to push the limits of software language
theory, but to design a language that fits some constraints of what
we've learned about software applications and patterns people use/
abuse. These are all great, and I might like some of these almost as
much as Objective-C. Add in AOP features and I feel you've surpassed
it. (see, I'm not just a fanboi ;-)
In Scala I've certainly used a lot of functional thinking to
implement designs:
- pattern-matching to select the appropriate behaviour in response
to an event
- use of a function-pointers and closures to define an event handler
(no clunky xxxListener interface needed)
- map and flatmap to implement component models in terms of other
collections
- higher-order functions
Yes - I do like how you can think O-O AND functionally in a mixed
environment in Scala. that's what I found with Objective-C as well,
but you can go typeless in Objective-C (the id pointer type with
selectors) which is harder in Java/Scala/etc. But that could all be
implemented.
plus actors, though these aren't strictly speaking unique to
functional programming.
Yeah - I like this paradigm - we're looking at this for noop too. (oh,
ironically, we've implemented most of the early interpreter for noop
in scala)
The point is that "best OO" still leaves a lot of other things to be
good at, it's by no means the end of the game! It would therefore
be a real shame if OO were the only realistic language-supported
option available to me (i.e. C++/objective C), and that has now
become the sad but unavoidable truth for iGadget development.
Also agreed. Yeah - my point, as I said, is not to defend apple's
decision - it was to object to the kinds of statements people are
reflexively making about Objective-C because of (what I perceive to
be) fundamentally emotional reactions to Apple's stupidity, being
taken out on a language that has, frankly, produced high-quality
software for over two decades. One might argue that it only recently
changed because it didn't much need to - it already had mixed OO/
functional capabilities, a dynamic runtime with typeless/typed
options, a really workable set of base-classes and application classes
with a nice GUI builder to assemble them (without generating code to
be maintained), AOP capabilities, and, with 2.0, garbage collection.
Most of the actual arguments are stylistic. It's a thoroughly modern
language in most respects... it's its style and position in history
and industry and fashions in software development which have relegated
it to perceptions of antiquity. I would argue that most of Java's
"evolution" was "fixing what had been broken from the beginning" so we
finally get somewhere workable. And since I was in Javasoft around
the time of Java 1.1, I have some credibility on that point.
Being one of the "best OO programmers" a person knows is a worthy
achievement and something to be quite rightly proud of. Then again,
so is "best COBOL developer".
This is absolutely true, and I've met some damned great assembly and
Forth programmers whom I have nothing but respect for.
cheers,
Christian.
--
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.
