Re: Classpath future?

[Mark Wielaard]

Hi,

On Thu, Jul 12, 2001 at 09:35:25AM -0400, Etienne M. Gagnon wrote:
> I have some questions about the future of Classpath.  Maybe this is not 
> the right place to ask such questions (and get answers); please point me 
> in the right direction if so.
I think this is the right list. Only for the licensing issues you might
want to contact address@hidden or address@hidden directly.

I think the impression is correct but I don't think this is a very
concious decision. My personal goal is to get a complete and free
standard java class library, but sadly I don't have enough time to
track all the free JVMs that are out there. Currently I am only trying
out my changes under gcj because I have a working setup for that
environment. (I should probably try to get more JVMs working.)
My personal impression is that more and more free JVMs are popping up
and more and more of them use Classpath (SableVM, Kissme, ORP, japhar,
gcj, ...) It is just that the gcj hackers seem to be the most active
ones and they contribute the most back to Classpath. But other JVM makers
do use and contribute code also and Classpath is happy to accept it.

I am not really qualified to give my opinion on the JNI/CNI issue.
I didn't study any of them very deeply. CNI just looks much more
comprehensible then JNI. You are right that we should try to use
as little as possible 'native' code in the core libraries. I don't
think anybody wants to have different native implementations, it just
so happens that the gcj hackers are the most active again and they do
have working code written in CNI and they really want to have their
key parts implemented that way. If you tell us that implementing a
JNI wrapper over CNI is impossible that is sad, because that would be
the most simple way to share our implementation with libgcj. I know
that other people believed that they code do either source code
transformation (through a perl script),by writing a wrapper library
or by using just a small subset of either CNI or JNI.
If you tell us this is not possible please enlighten us and we will
listen to your arguments. The current thinking is (as I understand
corretly what Brian Jones is doing) to have both a jni and a cni
directory that contains native code implementations. Maybe if we
keep the native part as small as possible this won't be such a
maintenance task, but I think only time will tell.

I don't think there is any reason to fork classpath because I think
you will find the Classpath hackers more then willing to listen to
you. Yes, we might not have time to test every free VM out there,
and yes, we might sometimes make technical decissions that are not
ideal for all free VMs out there. But if we had enough time we would
do it all perfectly :) Please help us make Classpath as usefull for
your project as it is already for some other projects. We will listen!

A last note about the licensing issues that you raised.
The GPL+exception clause is a strategic decision, there are already
proprietary implementations that do a good job. To get the most
help from outside we decided to use the current license which protects
the users and the GNU project as much as possible and get as much
help from outsiders as possible. The LGPL is not practical if we need
the help from embedded device makers because they will not accept it.
Please indicate what you think could happen to your code that would be
bad for you or the end users if it was licensed under the GPL+exception
clause that would not happen if it was licensed under the LGPL.
Also note that if you assign your copyright to the FSF you can still
distribute your own code under your own license. I think it is a
good idea to assign code to the FSF for this project since that way
the FSF can better defend it in court and they can easier make any
license changes in the future when needed.
(Please write to address@hidden or address@hidden directly to get
more precise information on copyright assignment and what guarantees
you get for how the FSF handles the code you assign and the rights
that you keep/get back.)

Thanks for your email. It is nice to know that some people find our
work so important that they care about the direction the project
is taking.

Cheers,

Mark
-- 
Stuff to read:
    <http://www.toad.com/gnu/whatswrong.html>
  What's Wrong with Copy Protection, by John Gilmore