Re: Importing tmux into base

[Matthew Mondor <mm_lists%pulsar-zone.net@localhost>]

On Sat, 12 Feb 2011 19:07:13 -0500
Christopher Berardi <cberardi%natoufa.com@localhost> wrote:

> On Fri, Feb 11, 2011 at 10:50:40AM -0500, der Mouse wrote:
> > > We could strip the base install down to something that even mouse
> > > would run -current.  :-)
> > 
> > Heh.  First, get rid of GPLv3.  (I've tried to read it and failed to
> > understand it, to the point where others have pointed out aspects of it
> > I completely missed.  That makes it unusable for me.  Even if TNF has
> > done better than I on that score, I would be the one on the putative
> > copyright-infringement hook, so that doesn't matter.)
> 
> I actually don't understand why there isn't more effort on removing all GPL 
> code from the base by finding existing alternatives or writing it ourselves if
> need be? 

You're welcome to help :)

I can see several factors:

- Lack of resources (small community, no salaried coders, no bounties
  other than what GSOC provides to a few students per year)
- Resistence to change and to new ideas by old folks
- Resistence to "old" tools (i.e. CVS) and ridgid quality expectations
  including a code review process, test unit code, documentation by new
  developers
- Small markets, with companies benefiting from BSD code not always able
  to contribute back (donations or code) and that's allright with
  MITL/BSDL (I won't mention the huge BSD-based vendors that don't
  directly contribute :)
- Status-quo "just working" and good enough for most users, and for
  embedded vendors the GPL code is isolated out enough; a fully BSD
  system is not needed by every BSD user (also look at the high number
  of users of GPL software such as Linux, to which this matters little)
- The GPL/LGPL licenses have so much momentum that most redundant tool
  implementations by third parties are also released under those
  licenses

I can't say that it never happens though, OpenBSD has some open-foo
projects which are cleanroom implementations of existing software,
NetBSD already has a pax tar frontend instead of gnutar; some BSD users
are slowly moving towards PCC or LLVM/Clang...  and a fair number of
base userland utilities are under MIT/BSD, but indeed a full
multipurpose BSDL system is probably not at the door.
-- 
Matt