Re: elementary OS

[David Chisnall]

On 11 Feb 2014, at 11:25, Rogelio Serrano <rogelio.serrano@gmail.com> wrote:

> I don't really care about current closed source companies.

That's a mistake, because in the real world there is no such thing as a 
closed-source company and an open-source company.  Microsoft has contributed 
code to FreeBSD, Red Hat has some proprietary products.  The question should 
not be 'does your company have a long-term philosophy to software licensing 
that I agree with', it's 'can we usefully cooperate to improve the Free 
Software ecosystem'.  

To me, one of the core philosophies of open source is the observation that 
cooperation can produce better results than competition.  

> I'm learning how to engineer compilers precisely because of LLVM. C++ makes 
> me want to pull my hair out by the roots. Slowly.

If you think C++ is bad, take a look at what happens when you try to write a 
compiler in C.  It's even less pretty.

> They do it because they can. Until we can gain more power by hook or by crook 
> that will always be the case. A machiavellian approach is needed.

Or we could try enlightened self-interest.  Companies contribute to FreeBSD 
because it's cheaper for them to do so than to maintain a private fork.  Over 
the last couple of years, we've had some big contributions from companies that 
had maintained private forks for a long time, but found that other people were 
implementing the features that they had in incompatible ways and merging was 
costing them a lot of developer time.  In contrast, pushing stuff upstream that 
isn't their core competitive advantage  (which, in most cases, is not even the 
software, it's the integration of the software with hardware and the 
surrounding ecosystem) costs them almost nothing.

No Machiavellian approach is required.  Companies are increasingly coming 
around to the idea that if they license software under a proprietary license, 
they're setting themselves up for a large long-term cost.  If they pay someone 
to improve software with a more permissive license, then they are not locked 
in.  They choose open source because it reduces their risk and potential 
long-term costs.  They avoid GPLv3 because their lawyers tell them that it 
exposes them to greater risk.

If you really want a Machiavellian approach, how about noticing that Apple has 
contributed a huge amount of code to Objective-C support in Clang (including 
the static analyser), which directly benefits GNUstep.  Wouldn't it be great if 
we could make more companies invest in things that benefitted open source 
versions of their proprietary products?

> They can afford it why not. If there is no company willing to use gplv3 I'll 
> make one. Gplv3 software is a strategic resource in my opinion.
> 
> Most companies find me unemployable so that's not a problem.

Good luck with that.  And then try selling your code in a marketplace that has 
decided that more permissive is better and views GPL'd code in the same light 
that the FSF views closed proprietary code.  I'd agree that the GPLv3 is a 
strategic resource.  It has been great for the FreeBSD project and for LLVM.  
Both projects have had a massive increase in both volunteer contributors and 
commercial contributions as a direct result of it.  

The fundamental difference between the BSDL and GPL philosophies is that the 
BSDL aims to make writing open source software easier, and accepts that the 
side effect of this is that writing proprietary software also becomes easier, 
whereas the GPL aims to make writing proprietary software harder and accepts 
that the side effect is that writing Free Software becomes harder.  

Personally, I believe that the best way to move beyond a world where 
proprietary software is the norm (as we've now done in a lot of markets) is to 
make writing Free Software as easy as possible, because single-vendor 
off-the-shelf software just can't compete in a world where people are 
cooperating to build exactly what they need.  Any impediment to this 
cooperation harms the ecosystem and promotes proprietary software.

David

-- Sent from my IBM 1620