Re: Plans for 2.8 - GNOME Managed Language Services?

[John Luke <jluke users sourceforge net>]

Hello,
I should probably remain a quiet bystander but I couldn't help it.

On Sat, 2004-03-27 at 12:03 -0500, Havoc Pennington wrote:

> On Sat, 2004-03-27 at 01:47, Ryan McDougall wrote:
> > MY understanding that all ECMA bits are unencumbered insofar as they
> > *must* be licensed under RAND terms. Is my understanding incorrect?
> 
> One problem is that RAND is still GPL-incompatible. That's why you can't
> ship MP3 or MPEG4 codecs licensed under the GPL, even if you buy a
> patent license.


RAND alone might be GPL incompatable, but Miguel has stated many times
that the ECMA submission was also royalty free. It was also reported in
one of the Oreillynet blogs that they may have an announcement after
doing a formal review to make this more certain and legally safe.

> > If we choose a open source JVM, then we get no language other than Java
> > and get Novell mad at us.
> 
> I don't know that's true; Novell had a Java thing back in the day:
> http://developer.novell.com/research/appnotes/1998/january/01/05.htm
> 
> But if true, it probably leaves us back at C/C++/Python.
> 
> >  If we choose a ECMA CLR standard VM and ECMA
> > CLI standard object system, we can implement a library stack in any
> > compliant language, and use it from any compliant language, including C#
> > or Java. How does that "lose support for GNOME"? If the objects are
> > highly interchangeable between languages, what does it matter what
> > language they are written in?? 
> 
> You're thinking in purely technical terms. It's unclear we're trying to
> compromise here between some people who want the positive attributes of
> C# and some who want the positive attributes of Java. Rather, some
> people want to avoid the negative legal/political attributes of C# and
> some want to avoid the negative legal/political attributes of Java. So
> saying "allow both" isn't a workable compromise, two sets of negatives
> are twice as bad, not twice as good.
> 
> Similarly, if we prove Java is bad we rule out Java, we don't rule in
> Mono. If we prove Mono is bad we rule out Mono, don't rule in Java. If
> they're both bad then it's C/C++/Python.

So are we past the point where we are not deciding this on technical
terms?  I think knowing that the actual reason mono can not be used is
legal/political helps this debate greatly.  I don't think anyone has
explicitly said that until now.  That way it is very simple for Novell
to know what they need to prove in order to allow for adoption of mono
in the core of GNOME.

As far as I am concerned (maybe I am biased), Java and the JVM offer no
advantages over mono.  But mono has several including: language interop
(managed to unmanaged and also the reverse), complete and mature open
source implementation based on two published standards, a minimum of
RAND license granted on patents covering that standard, a large and
vibrant community behind it that already write GNOME apps and would be
willing to make accomodations for the needs of GNOME.

Sticking with C would be workable, although not as desirable as mono, as
Gtk# already works with that very well. Allowing Python into the core
might not work as well for working with other languages.