[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: elementary OS
From: |
David Chisnall |
Subject: |
Re: elementary OS |
Date: |
Tue, 11 Feb 2014 11:04:55 +0000 |
On 10 Feb 2014, at 21:30, Gerold Rupprecht <geroldr@bluewin.ch> wrote:
> The newer syntax in Objective-C require some hacking on GCC, the
> compiler collection. David Chisnell is very capable, but has been
> working mostly on Clang. Getting similar features into GCC would be a
> big plus. David, I would love to hear your thoughts if this would be a
> fruitful endeavor to broaden the appeal of Etoilé?
GCC support for Objective-C is dead, and GCC support for ARM is likely to
languish. I talk quite regularly to ARM's compiler group. They have customers
in two categories:
- Won't use GCC because of the license (GPLv3 means to a lot of companies
'don't let this code in the door')
- Will use whatever compiler ARM recommends
As such, ARM is focussing entirely on LLVM and encouraging their partners to do
the same. The two big mobile operating systems, iOS and Android, are backed by
the two companies that employ the majority of LLVM developers.
The license is, to be honest, also an impediment for GNUstep. I'm seeing
increasingly that commercial entities have two strategies with open source:
- If it's copyleft, fork it and don't tell anyone. Obfuscate your binaries a
little bit and hope no one sues. This generally works, because most open
source projects either don't have the resources to sue, don't care, or don't
want to frighten off other companies by being seen as litigious.
- If it's permissively licensed, fork it and upstream anything that doesn't
give you a competitive advantage, so that your merge costs are lower (and so
other people fix your bugs).
There are two or three companies that I strongly suspect of following the first
approach with GNUstep and I think we missed an opportunity with them, as each
is independently putting more manpower into their private fork than we are on
upstream.
Having the tools under GPLv3 is also a problem, because shipping working
programs requires shipping many of the tools (at least defaults and gpbs) and
that puts companies off. It's no accident that most of the biggest users of
GNUstep use it for in-house development and not distributing the result.
I find it increasingly difficult to be motivated to work on a project where I
most likely can't use the results commercially because it has a license that is
not permitted by most of the companies that I work with. I am not the only one
who has this problem: Nicolas (who used to be a very active contributor) is now
writing an entirely new GUI toolkit on his employer's time because they don't
want copyleft software touching their products.
David
--
This email complies with ISO 3103
- Re: elementary OS, (continued)
- Re: elementary OS, Michael P. Soulier, 2014/02/10
- Re: elementary OS, Riccardo Mottola, 2014/02/10
- Re: elementary OS, Michael P. Soulier, 2014/02/10
- Re: elementary OS, Dr. H. Nikolaus Schaller, 2014/02/10
- Re: elementary OS, Gerold Rupprecht, 2014/02/11
- Re: elementary OS, Gregory Casamento, 2014/02/11
- RE: elementary OS, a b, 2014/02/10
- Re: elementary OS, Riccardo Mottola, 2014/02/11
- Message not available
- Fwd: Re: elementary OS, Rogelio Serrano, 2014/02/11
- Re: elementary OS, Gerold Rupprecht, 2014/02/10
- Re: elementary OS,
David Chisnall <=
- Re: elementary OS, Rogelio Serrano, 2014/02/11
- Re: elementary OS, David Chisnall, 2014/02/11
- Re: elementary OS, Pirmin Braun, 2014/02/11
- Re: elementary OS, David Chisnall, 2014/02/11
- Re: elementary OS, Pirmin Braun, 2014/02/11
- Re: elementary OS, Liam Proven, 2014/02/11
- Re: elementary OS, Riccardo Mottola, 2014/02/10
Re: elementary OS, Riccardo Mottola, 2014/02/09