On Feb 1, 2007, at 10:45 PM, Stephen Lau wrote:
On Thu, Feb 01, 2007 at 08:40:00PM -0800, Scott Tracy wrote:
Having the source ready for publishing is not a requirement for
setting up a project. I'm not looking for an endless debate here
but there are many examples of Open Solaris Projects that started
without the source. iSNS is just one such project, in fact it
only has prototype code at the moment, but it started with
nothing. The idea is to get the community educated, involved and
yes, eventually publish the source so a developer can DO something
with it, but it's not a barrier to entry as long as the intention
is not malformed into a marketing ploy or other sleezy engagement.
1) I voiced earlier that I was disappointed at previous projects that
didn't have source available either, and that I was going to be a jerk
about it from now on.
I don't want that to be a gating factor to get community involvement
and create a beachhead for a project. In fact I can see where the
start of a project is a design doc and both folks in and outside Sun
do joint development to get a project started/completed.
2) The difference between Honeycomb and iSNS is that iSNS is
targeting a
consolidation (NWS) that already publishes source - so even if the
iSNS
developers never make their own source tarballs available (which I
hope
won't happen) - we at least have a reasonable assurance that it
will be
open in the end by way of the NWS consolidation. On the other
hand, the
best I've seen Honeycomb offer is "we will look at opening up
Honeycomb". That's just not strong enough of an assurance for me.
How would anyone outside of sun know this? And why does it matter
anyway. Refer to above, it cannot be a gate to start a project by
having source first. That's not the objective. If that's the case,
then how will the community do joint projects or have influence on
ones in progress? The idea is to be open through-out the process.
Maybe I'm missing the point. What problem are you solving by gating
folks without code from starting a project? Are there too many of
these that have failed or something?
The SDK layer is not cast in stone. XAM is in SNIA now, but is
nowhere near complete. This is a chance for developers to take a
look at the interfaces, figure out what they want and influence
the product (and potentially the industry). I'd say this is open
development.
Sure, so why does it need a project?
The only thing a project offers over communities are SCM repositories.
Clearly, Honeycomb does not yet have source - that's fine. Peter has
already initiated discussion the Appliances community, let it live
there
until they have source ready - at which point they can request a
project
to host their source.
So, let's talk through this so I can understand. How would you work
on a design doc with the community? Walk through the normal SDF
process on something you want to develop from start to finish in the
open.
I would give my vote to start the project provided the team starts
with a reasonable set of documents and SDK to make my mouth water
as long as I have the details to use it. Something like:
* Design Documents on the source
* SDK
* Example application on using the SDK (or Whitepaper)
That's fine... that's why we have these public proposals so people can
get +1 or -1 opinions from people. :)
Yep.
/Scott
_______________________________________________
opensolaris-discuss mailing list
[email protected]
