ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jason Hunter <>
Subject Re: Ant vs. other "ant"s (was: Re: Whoa Bessie...)
Date Fri, 22 Dec 2000 02:50:21 GMT
Diane Holt wrote:
> Hi James,
> I think I understand how you feel, and how things managed to get confused
> and very nearly completely out-of-control -- but I also think the two most
> divisive things that happened were 1) people had gotten used to Ant being
> an OpenSource project, with all that implies, and 2) you *seemed* to be
> trying to un-OpenSource it by saying you had returned to take it back over
> and (I'm paraphrasing) "make it be what it should've been from the start
> and would have been but for your having to leave it behind for awhile".

An open source code base can't be made "un-OpenSource".  Under the code
license (you should read it if you haven't) everyone is free to use the
code, modify the code, and redistribute their modifications.  That last
part means everyone is free to "fork".

I'm not saying he should, but it's important to realize James is legally
able to take Ant and derive a new work, and he can do what he wants with
that work.  You and I and everyone else have the same rights.  

Yet while we all have that right, there is (and should be) great social
pressure to build consensus.  Unnecessary forking fragments the
community.  It cuts developer resources, confuses users, and is
something to be avoided where possible.

On the other hand, forking is sometimes the best thing for a codebase. 
It can allow experimentation that the main codebase couldn't tolerate
(see Samba's "Samba TNG" fork, endorsed by Samba's original creator). 
It can be a great way for those with new ideas to prove their ideas,
without needing to build "consensus".  And forking can be a great way
for those with bad ideas burn themselves out on their own project,
without needing to build consenus.  :-)

I can't predict the future here.  I think it's best to try for
consensus.  Maybe that's not possible.  It sounds like people may have
different goals:  I see one wants to optimize for building Java programs
and one wants a generic approach.  If projects must diverge, that's when
things get interesting.

Only *one* of the forks should have the name "Ant".  If the "fork" were
done outside Apache it'd be clear that the ASF project was the one named
Ant (see the license; others are legally disallowed from using the name
"Ant").  But... when the forks are done inside Apache, then it's less
clear.  I suspect James feels he has most right to the fork named
"Ant".  I don't think anyone (James included) would say he has the right
to kill any other forks or tell any developers which fork they
contribute to.

My previous post addresses my thoughts regarding naming.


View raw message