Re: [Gnu-arch-users] Re: arch roadmap 1 (and "what's tom up to")

[James Blackwell]

In lists.arch.users, you wrote:
> address@hidden (James Blackwell) writes:
>> Let's hypothesize for a moment. Hypothetically speaking, Tom goes
>> bonkers and decides that from now on, instead of using english, arch
>> would use piglatin. This most unpopular decision would result in the
>> other significant developers to start following a different tree. Tom
>> would be left alone.
>
> Sure, but there's a lot of hysteresis -- most tla users already trust
> Tom, and are used to him, and so won't abandon his tree without a fairly
> strong incentive (at least quickly; if someone else's version of tla
> were just "slightly" better, there might be a gradual drift of users to
> it).

Yes. Thankfully there is hysteresis. If hysteresis in free software
development didn't work, I wonder if free software could work at all.
Everybody would constantly be forking everything. This gives the project
leader (Tom, in this case) a bit of wiggle room because he can say "just
trust me". As long as (most of) his decisions work out, he's good to go.

But at this point insert your favorite cliche about people disliking
change. The further ahead any given idea is, the harder he has to
work to promote it and gain consensus.

> That means that Tom (like Linus) has a fair amount of wiggle room, and
> _can_ make decisions that are unpopular in the short-term without losing
> his users.  This is a very good thing because sometimes such actions are
> necessary.

LOL. We used the same cliche. :) Yeah, I obviously agree.


> [Of course forks are not always so clear-cut anyway, if the maintainers
> of different trees are all merging from each other (e.g., the different
> BSD forks -- though judging from the abysmal OpenBSD benchmark numbers I
> saw recently, they're not doing it enough!).]

I think that forks usually happen for one of two reasons: either
mainline development stalls (or at least goes too slowly for comfort),
or the leader makes a series of decisions in a row without getting
others to buy into it first. 

Stephen, this would be a good time for you to jump in. I know the
emacs/xemacs fork was a pretty big one, but that's about _all_ I know.
Would you have any illuminating anecdoctal evidence? 

-- 
James Blackwell          Try something fun: For the next 24 hours, give
Smile more!              each person you meet a compliment!

GnuPG (ID 06357400) AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400