> Hello there,
> I just made a first tour of the doc and examples. Did not try it yet.
> I'd like to know if anyone has some performance info/comparison.
> If JSCocoa is to be used as a main language for an App, or a
> consistent module, its code is interpreted, right? Then, its speed is
> like JS running in Webkit? Or better? Or what?
I have no benchmark figures.
JSCocoa is just a few custom JavascriptCore objects running in a
standard Javascript context, like Safari. Every time you access an
ObjC object, there is some data conversion (boxing and unboxing)
between ObjC and Javascript. This means the less often you use ObjC,
the closer you'll match Webkit speed.
Regarding the intrepreted code - it's interpreted if you're running
Safari 3. In Safari 4, code is compiled at runtime, just like in
Safari. You'll never be faster than JavascriptCore, only a bit slower.
-Patrick
> Now how fast is compiled JS vs. compiled ObjC is not a JSCocoa issue,
> fore sure.
> Nevertheless I'd like to figure it out before jumping in.
The only way you're going to find this out, is to try it out.
Each app has it's own performance problems. The best thing to do is
just to come up with a proof of concept of what you think would be the
bottleneck, and judge for yourself if it's fast enough.
-gus
--
August 'Gus' Mueller
Flying Meat Inc.
http://flyingmeat.com/