[opensolaris-summit] My comments (very subjective) on proposed Summit topics

[Alan DuBoff]

On Wed, 26 Sep 2007, Richard Lowe wrote:

> We don't lighten the process, we certainly don't drop any quality
> control aspects thereof, we just make it work without the need for
> someone on SWAN to hand-hold, go try and find where a bug ended up,
> etc, etc.  I'm not saying that this shouldn't be the way *everyone*
> (to include new hires at Sun) works with our gates at first.  I'm
> saying there needs to be an alternative.  The sponsor process doesn't
> scale, and was never intended to.

For better or worse, Sun is moving to get outside the SWAN, it's just 
slow, but as I said in a previous message, I can only fault them on the 
time it's taking.

There is no having your cake and eating it too, if you want to keep the 
processes in place and not lighten up on anything, it will continue to 
require a sponsor. The process is really too complicated for the average 
person to follow without help, IMO.

However, testing is the most difficult piece with device drivers for 
instance. Ok, you don't want to lighten the process, it will continue to 
take a long time to putback each driver and work through the issues and 
fix them to be able to meet the rigorous testing that is currently in 
place.

I feel the testing needs to be evaluated in some cases. For instance, 
testing a NIC driver to be loaded and unloaded thousands of times will 
probably just tell you if there is a memory leak. There could be better 
ways to test that, IMO. Loading and unloading a driver is not a common 
practice on most systems.

>> Can you elaborate on what you would do different in regard
>> to being more open in the development of the current sources?
>
> We get to the point where external developers have equal access to the
> tools and processes, and useful things can be done without latent
> intermediaries.  I really don't care what developers inside Sun (or
> outside Sun) choose to do.  Only that they have the ability to
> actually do it themselves, and that the decisions made about changes
> made to opensolaris gates are taken by opensolaris, in places
> opensolaris can see, rather than by Sun alone.

Right, when you can do a putback to OpenSolaris, and Sun can determine if 
they want to accept that into Solaris, then we will have reached that 
level. FWIW, we are moving there, it's just moving there slowly.

--

Alan DuBoff - Solaris x86 IHV/OEM Group