[Discuss] licensing: who freakin cares?

[Robert Krawitz]

On Sun, 10 Apr 2016 14:08:01 -0400, Rich Pieri wrote:
> On 4/10/2016 5:19 PM, jc at trillian.mit.edu wrote:
>> The open-source work I've been involved in has rarely acted this way.
>> Part of the reason is that if the leaders try it, people just quietly
>> drop off the team and start working on something else.  Or they  fork
>> the  project  and do the needed work themselves (leading to the usual
>> hassles if they try to merge it back into the main package).
>
> This is the promise of free and open source software. When it works it
> works really well, but when it fails it fails worse than any by the
> dollars MBA. The GCC cadre is my go-to example of such failure. Lack of
> innovation (GCC falling behind commercial and open source compilers'
> performance), failure to keep up with standards (GCC's C and C++
> compilers are way behind current standards for these languages),
> rejection of contributions from outside sources (the rejection of Clang
> integration), and low general code quality (everything Linus Torvalds
> ranted about GCC 4.9).

That's a project management/governance issue independent of choice of
license model.  GCC has had (more than) its share of NIH, if you
recall; there was a project named egcs that forked gcc, precisely
because gcc wasn't being maintained.  Eventually by mutual agreement
egcs became the new gcc mainline.  Sad to see that that issue has come
up again.  But this kind of thing is hardly specific to the FOSS
world, much less the GPL world.  It's not uncommon for products to be
orphaned; at least in the FOSS world, it's possible to do something
about it.  Look at Revolv for an example of what can happen with a
closed product.

But the Linux kernel is also GPL-licensed, and it doesn't suffer from
those problems.

> And as a second example, one that I was briefly involved with: Claws
> Mail. One of their core developers stated that losing mail is an
> acceptable tradeoff for performance and the others did not dissent. To
> say that I was horrified by this is an understatement. Losing mail is
> unacceptable. Period.

Agreed, but it's hardly the only consumer-grade product that has gone
out with flaws of that nature.

> You generally don't see this kind of management behavior in aerospace
> and heavy industry. The stakes are far too high. After all, you can't
> fix it in production when your product is in an Airbus A350 cruising at
> Mach 0.85 at 40,000 feet above the Atlantic Ocean.

No, but that's also not what most consumer products (of any license,
proprietary or open) are qualified for.

> And "we'll fix it in production" turned out not to work for Facebook,
> either. Their old "move fast, break things" philosophy did not sit will
> with their customers. Paying advertisers and developers didn't like it
> when their ads and games weren't being served to users because chunks of
> the infrastructure were broken. So, when push came to shove, Facebook
> changed their philosophy. It's now "move fast, stable infra" because
> they don't get paid when production breaks.

Which again, is entirely independent of license model.
-- 
Robert Krawitz                                     <rlk at alum.mit.edu>

***  MIT Engineers   A Proud Tradition   http://mitathletics.com  ***
Member of the League for Programming Freedom  --  http://ProgFree.org
Project lead for Gutenprint   --    http://gimp-print.sourceforge.net

"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton