Re: [Pan-users] Weird bug with gcc-5.2's libstdc++ (sort of solved)

[Duncan]

walt posted on Wed, 09 Sep 2015 05:06:17 -0700 as excerpted:

> Very interesting.  I never considered licensing as a driver for llvm.
> Now I'm wondering about mir (from ubuntu) and wayland as replacements
> for X.  I'd heard that those projects are driven by the need for
> something simpler and faster than X, but now you've got me wondering
> about licensing, too.

Wayland, as the intended X successor, is deliberately licensed similarly, 
MIT.

Mir... is an ubuntu-specific project that isn't playing well with the 
rest of the free/libre and open source community, so I'm not sure on it.

I do know that one of the ways in general that Ubuntu isn't playing well 
with the rest of the community is that for their projects they typically 
require either direct copyright assignment or agreements that let Ubuntu 
relicense to proprietary, but as the MIT license already allows 
proprietary relicensing, I'm not sure of what they did with Mir.  Maybe 
it's straight MIT and the not-playing-well related to it is simply 
refusal to cooperate, not licensing, but I don't know that as I've not 
investigated (and no gentoo dev has apparently been interested enough in 
it to add it to the gentoo tree, so I can't simply look it up as I can 
with most projects I'd be interested in license status on, either).

But I did just look up mir on wikipedia, and it says the license is 
GPLv3.  That again is with Canonical requiring the right to take it 
proprietary, which would *DEFINITELY* rub the community the wrong way, 
because it would let only Ubuntu take it proprietary, angering the free 
as in freedom copyleft folks, the BSD/MIT nothing wrong with proprietary, 
but let everyone do it, folks, and the just don't rock the boat, the 
existing X solution is a standard reference implementation with the MIT 
license and any successor should be similarly if not identically 
licensed, folks, all three.

There's more on the wikipedia page.

https://en.wikipedia.org/wiki/Mir_%28software%29

(As should be obvious, I'm rather interested in licensing issues, since I 
believe licensing helps keep freedomware free, ensuring freedom not just 
for me, but for every user and developer of that software and derivatives 
thereof.  I'm not always opposed to dual licensing, and indeed, supported 
Trolltech/Qt with its former dual licensed model, in part due to the Free 
Qt Foundation (name?), etc.  But IMO, as a standard-target, the non-copy-
lefted MIT license is entirely appropriate for X and successors, 
particularly so since X itself has the long and strong history it does.  
So along with the example of Internet protocols and reference 
implementations, X and its successors, due to their universal standards-
target nature, are exceptions to my default strong copyleft stance, and I 
find the MIT license entirely appropriate.  In fact, even more 
appropriate for them than a copyleft license such as GPLv3, particularly 
when some company is trying to reserve the right to take the otherwise 
standard proprietary to itself, while forcing others to copyleft their 
changes as well as grant that company the exclusive right to take them 
proprietary if they want those changes integrated, something I have less 
problem with in the non-standard context, for example, the qt/trolltech 
case.)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman