From: Austin Ziegler <austin@...>
Date: 2002-09-09T23:01:57+09:00
Subject: Re: JRuby (was Re: OS-independent build of ruby)

On Mon, 9 Sep 2002 18:36:41 +0900, Anders Bengtsson wrote:
> Austin Ziegler wrote:
> 
>> However, as someone who uses Ruby and might consider JRuby, it
>> does matter to me. As such, it's definitely my right and
>> responsibility to comment. It's also my right and responsibility
>> to raise this issue in a forum where the original copyright
>> holders -- who appear not to have been consulted about the choice
>> of only one of the two available licences -- can be notified.
> I've seen smoother entry to a user community than calling the
> software's licence "nasty". ;-)

Fair enough. In my defence, though, I will note that I was
considering the limitation of the licence from the two that Ruby has
chosen to the one that is (IMO & E) least useful for the chosen
implementation framework to be seemingly nasty. Again, it's not so
much the licence being nasty as the apparent violation of the
*spirit* behind Ruby.

>> I think, then, that there's a problem. My understanding is that
>> when a program is dual-licensed, you have the option of accepting
>> it under either of the licences available. However, modified
>> versions *must* be released under both (or all) existing
>> licences. I've put this question to Lawrence Rosen
>> (lrosen@rosenlaw.com) who writes for one of LJ or LM. I haven't
>> given details except the fact that the original project is under
>> GNU GPL and/or artistic-style and the forked project is only
>> available under GNU GPL.
> Once one has chosen a licence, that licence applies to both
> useing, modifying and distributing, as far as I understand. But
> I'm not really interested in any legal hair-splitting. Our
> intentions are simply to find the licencing that works best. As
> you so bluntly pointed out, GPL-only doesn't work well for JRuby.

I think it's an interesting question, and I'll definitely be
forwarding on any answer that I get to the question -- but I think
that the resolution is to give JRuby a better licence. As I've said,
I think that the LGPL is a better choice than the GPL, especially
for working in a Java environment (or possibly in a .NET environment
if it's ported later).

>> I am curious, though: why didn't the main contributors want to
>> use Ruby's licences (both of them)?
> From what I've heard, the problem was mostly that a few details in
> the Ruby licence don't work well with a Java program. But the main
> reason is probably that the Ruby syntax parser that was used as
> the start of the JRuby project was GPLed, so the easiest thing was
> to just continue with that. Now we're in the middle of changing
> licences, but these things take time.

I don't understand the former. I can see nothing in there which
is incompatible with Java in any way, whereas the GPL is wholly
incompatible with fully dynamic environments (in part because of
Stallman's interpretation of dynamic linking and linking as being
identical).

The latter make sense, but it still seems strange to me that a bit
more effort wasn't made to make sure that item was available under a
more liberal licence (even the LGPL) in the first place, which would
have freed up the choices for the rest of the developers. Of course,
such effort could have been made and I simply don't know about it.
(Then again, I'm always baffled by people who do Java stuff under
the GPL, when it's clearly inappropriate because *everything* is
dynamically linked and therefore ripe for instant and continuous
violations of the GPL.)

-austin
-- Austin Ziegler, austin@halostatue.ca on 2002.09.09 at 09.39.28