JIT?
SmallSqueak
smallsqueak at rogers.com
Sat Dec 18 12:24:11 UTC 2004
Avi Bryant wrote:
>
>> When Java crossed some critical performance threshold it
>> took off, because it got taken seriously. Squeak is still waiting to
>> be taken seriously. Yes, it is used in many research contexts, but not
>> much elsewhere by anyone that is not a die-hard Smalltalk fan. There
>> is very little that things like Ruby offer that are not far better done
>> in Smalltalk, but they are out there and we are gathered around the
>> fire.
>
> You probably know this, but it seems worth pointing out since the
> context implies something else: Ruby is considerably slower than
> Squeak is. Ruby's success, and the degree to which it is taken
> seriously, has nothing to do with performance, and everything to do
> with a very savvy combination of Smalltalk's core semantics and
> libraries with a syntax and environment that's familiar to typical
> *nix or Java developers. It also helps that Ruby's C extension API is
> the cleanest and simplest I've seen, which means that wrappers for
> many significant C libraries appeared pretty much overnight.
>
I don't think that's the main reason.
Ruby doesn't have something that Squeak has, a monolithic image
which has been continuosly polymerized for decades.
> *If* we want to make Squeak much more popular and much more widely
> used, there are no doubt lessons (or, heck, syntax and libraries) that
> we could take from Ruby. I'm personally not very convinced this is
> the goal we should be pursuing, however...
>
It is a real challenge to turn the monolithic image into modular
packages.
Even Alan and Dan endorse the vision of Modular Squeak.
Cheers,
SmallSqueak.
More information about the Squeak-dev
mailing list
|