[SF][VM] SF for VMs: Phoenix out of ashes?

Lex Spoon lex at cc.gatech.edu
Sun Feb 11 02:42:56 UTC 2001


Bert Freudenberg <bert at isg.cs.uni-magdeburg.de> wrote:
> On Thu, 8 Feb 2001, Raab, Andreas wrote:
> 
> > I believe that everyone who is involved in such a project for a long
> > enough time will have phases during which one is either too busy or
> > simply too tired of doing the necessary work.
> 
> Which is okay. But in a community project the work shouldn't have to stop
> for long just because one person is hindered. Further development should
> be possible, even if the next release has to wait until the maintainer
> finds the time for it. And "further development" does not just mean that
> everyone is hacking stuff on his own (this is what we're doing now), but
> that it is possible to aggregate single contributions into a common
> source tree (which must not necessarily be the "official" source tree,
> although it should not deviate too much).
> 

The general model of having a central maintainer and not having multiple
cooks in the kitchen is fine.  However, it hurts when releases are far
apart:

	1. Contributors duplicate a lot of work, both with each other and with
the central maintainer.

	2. The central maintainers get a lot of patches that don't patch
against their current code base.


Frequent releases are a way to let developers at large help out the
central maintainers.  Just consider the Squeak image and its update
mechanism....


-Lex





More information about the Squeak-dev mailing list