Compiler Source

[Maxim Kizub]

Edwin Foo wrote:
> 
> >PolyJ:
> >
> >http://www.pmg.lcs.mit.edu/polyj/
> 
> As a MIT student who has seen undergraduate students suffer under this
> compiler in our software engineering classes, _please_ don't use this one.
> =) It was the cause of no end of headaches and many late nights with
> coffee. If you look at http://www.pmg.lcs.mit.edu/polyj/bugs.html you'll
> see what I mean.

I don't think that bugs may be corrected if noone will use
compiler. The problems is not in bad test or something like this.
Peoples used to write very differently, even if use the
same language.
Compiler writers should test their products on
"external" projects, but, non-java code is hard to find
out ;-). The best case is when the compiler is written
with the full usage of extensions it offers, like
pizza or kiev.
If you like one compiler - just use it and send bug
reports to developers (even every hour, if you have
bugs offten ;-)). I have expirience, that amount
of bugs you can find is about from 5 to 10 bugs.
After that compiler is adopted to understand your style
of programming ;-)
Well, one bad thing can happen - developers do
not react quickly on bug-reports. Is this happened
with polyj and you?

> If polyj stabilizes and is less crash-prone it could well be a very cool
> compiler, although I can't say that I personally think that Java needs
> parameterized types either.

I've not found a good usage for parametrization
except of container classes. But for containers -
they are superior.

Regards
  Maxim Kizub