Okay, just a bit of nitpicking here, but this sort of clarification is 
shows that drawing analogies can be very tricky in this realm.

On Wed, 24 Oct 2007, Simon Phipps wrote:

> Apologies, catching up after three contented days of exploring
> national parks.
>
> On Oct 20, 2007, at 10:01, Stephen Lau wrote:
>
>> We are not like Ubuntu in that our projects and our code target
>> multiple distributions.  The work happening on Ubuntu's site
>> ostensibly is all targeted towards Ubuntu.  The work happening on
>> OpenSolaris.org targets multiple consolidations, multiple projects,
>> and multiple distributions.

A part of Ubuntu is their Launchpad development site, which both hosts 
Ubuntu development and packaging infrastructure and serves similar 
purposes for a number of seperate projects. The system is able to express 
relationships between "external" projects (even off-site) and the Ubuntu 
derived packages to attempt co-ordination of development processes.

>
> Except Ubuntu does result in multiple distributions - Edubuntu,
> Kubuntu, Gobuntu, xubuntu to name but four.

They are not seperate distributions in the sense that they are defined as 
different "metapackages" drawing from a single package pool, analogous to 
package metaclusters (and indeed similar in implementation). They do have 
quasi-seperate groups of people working in different directions but it 
seems to be mostly inside the same community.

>
> As for not drawing parallels; I agree to the extent that we should
> not actively model ourselves on another community. I believe that
> highlighting the mis-match of expectations in the cases where some
> people assert OpenSolaris is "just like the Linux kernel" as others
> have done is important though. We are at a turning-point in the
> OpenSolaris community, since it's clear (to me at least) that our
> initial assumption of a kernel-based community with many external
> distributions is no longer a good model. This is not least because
> (as Ian points out) it fails to deliver the easy ability for there to
> be a large pool of compatible applications.
>
> S.
>

Yes, looking at the problems exposed by other models is extremely helpful 
for avoiding these pitfalls.


-Albert

Reply via email to