Improving Java for Linux

Ian Brown inb at
Fri Nov 7 08:51:11 PST 1997

John D. Gwinner wrote:

> Again, not my original comment; Jim Pick said it.
> What I was talking about was a slight modification of this -- a 'subset'
> of Java if you will, which I think Jim was also thinking of.
> Why is that foolhardy?

The basic problem with this is that it makes things non-portable. If you do
not support exactly what Sun is producing, you end up either not being able
to run other people's code (because features are missing from your subset) or
your code won't run on any machine that doesn't run your VM (because you
implemented features missing from the Sun stuff).

> > Wait until ISO standardizes.
> On what?  If ISO standardizes on a language that is wholly owned and
> controlled by Sun/JavaSoft (whoever) then I've got big problems with that.

Why? Do you run MS Windows? If so, do you have a problem with the fact that
MS wholly owns and controls it?

> > If ISO standardizes, Sun will follow, and so should free-dom.
> Possibly.

Yes, this is more of an issue. If Sun doesn't follow the standard from ISO,
that might worry me. However, it is likely that ISO will end up standardizing
on something defined by Sun.

> Well, that's a major way right there!  If I can't say that my program is
> written IN Java, then that's a bit of a restriction, isn't it?

Did they say that you couldn't say that your program was written in Java?
I thought you were saying that you couldn't say that your program could run
Java code, which is rather different.

> It doesn't cost me anything to advertise that my application is written in
> C++.  What happens to me if I say that my native program can include Java
> programs?  Now I've violated a trademark, and you better believe that will
> hurt.

Which is different from the statement above about not being able to say your
program is written in Java.

> I'm NOT a Java expert, but who decides what the language consists of?
> Isn't it's Sun's right to extend or modify the basic VM and language, but
> no one else's?  Anyone that can change something with no one else having
> the right controls something -- and makes it proprietary.

Yes, that is true. It is not obvious that that is a totally bad thing. It has
the advantage that if I write a program in Java, it will run under any VM
that calls itself a Java VM. We don't end up having many different versions
of the language the way we do with C. I like the fact that I don't need lots
of odd '# ifdef <weird machine>' lines scattered throughout my programs.


More information about the kaffe mailing list