Re: For those interested in the Dreamcast

[Paul Mundt <lethal@linux-sh.org>]

On Mon, Jan 12, 2009 at 08:16:54AM +0000, Adrian McMenamin wrote:
> On Mon, 2009-01-12 at 16:54 +0900, Paul Mundt wrote:
> > On Sun, Jan 11, 2009 at 07:48:48PM +0000, Adrian McMenamin wrote:
> > > Veterans of the Dreamcast development community will recall we used to
> > > have a pretty good general mailing list, but it long ago fell into
> > > disuse.
> > > 
> > > I know the community is much smaller now than before but it would be
> > > good to have a general list, so I have set one up. Please consider
> > > subscribing
> > > 
> > > http://transpero.net/cgi-bin/mailman/listinfo/dreamcastdev
> > > 
> > Please explain how more fragmentation in this area is going to help
> > anything. It seems like you are aiming for one mailing list per dreamcast
> > developer.
> > 
> > Any OS specific stuff needs to go through the port-specific list, and
> > architecture specific issues can be cross-posted across those as well.
> > The only thing a dreamcast-specific list would be useful for would be
> > discussions of native development on bare metal that people don't see a
> > good fit for anywhere else. However, even that stuff is likely marginal
> > enough at this point that there is no reason to discourage such threads
> > from popping up on the OS-specific lists.
> > 
> > When you are already suffering from lack of activity due to your platform
> > being in its death throes, more mailing list fragmentation is definitely
> > not the answer.
> 
> If you don't want to subscribe that is up to you.
> 
> In the meantime this is not further fragmentation but the opposite - an
> attempt to (once again) have a single point where cross-OS issues could
> be discussed.
> 
You seem to have entirely missed the point about cross-posting being OK
for these situations today. The linux-sh list is certainly open to
cross-posting, regardless of whether people are subscribed or not, and
general architectural issues are certainly best discussed there.

The very definition of fragmentation in this context is when a small
marginal interest group runs off to start a separate forum for discussion
with the express purpose of only focusing on a fixed issue. In practice
this tends never to work in the long haul, so you either have people
subscribed to both places, or you miss out on the knowledge base
available on the higher volume list (including the people who designed
the hardware in the first place).

Encouraging discussion within your own interest group is great, but there
is no reason why this can not be done on the existing lists today without
trying to drive another wedge in to the development community. There are
simply not enough developers for a new list to be necessary.

The time to split off a list is when there is too much chatter about a
particular area of interest on a more general purpose list that it
actively becomes a distraction, and there are enough people and momentum
to carry it forward. Only in these situations can fragmenting be useful,
and these lists tend to fall in to disarray and die off completely once
that particular issue has been dealt with, or the momentum falls off
naturally. Premature fragmentation just means that you have small pockets
of people with similar interests split off in to totally separate groups
with very little communication between them.