First off,
Thankyou for responding to me/Alan/JimG/...
S Destika wrote:
Now all of what you said makes sense when you think of it from Sun's
PoV - all is well when it comes to Sun wanting to enforce control
over the whole OpenSolaris thing. This is exactly how a company will
drive it's product development.
If you cared to come out and peek into a general oss coder's PoV -
all these bureaucratic things make much less sense if any. And the
problem that I have with Sun's processes is more that they don't
create a sense of ownership in the OSS developer.
I'm very keen to find out how Sun's processes fail to create a sense
of ownership in you as an OSS developer -- would you care to elaborate?
Sun needs to discover this so that (Open)Solaris can be made better.
People aren't just
going to come and work if they have no motivation - motivation comes
significantly from owning and being able to drive something.
Agreed, wholeheartedly. And you won't get any disagreement on that
from any of the other Sun participants either.
This old school thought of Sun in head master's role and devs in
> student role just doesn't cut it.
That is not the attitude that I have observed amongst any of the
(Open)Solaris developers. If you've seen this in person, please let
me/JimG/Simon Phipps ([EMAIL PROTECTED]) know and that will be
addressed.
Linux succeeded not because of quality motivated people to enable
RedHat/SuSE do something under their control - it succeeded (and
succeeds) because of lots of motivated people who trust Linus (Linus
refuses to be employed by any of the vendors - exactly because of
this reason, he doesn't want to break the trust.)
And you know what? we are _all_ the richer for it.
Other lovely thing about Linux is that you are free to do what you
like - and not be driven by what Linus or Andrew think. Ok, so you
didn't like GIT for some reason, take Mercurial or do your own SCM
and move along - work continues to happen on the kernel at the same
pace.
This is already taking place both within and without Sun. Casper Dik
(on the CAB) seems to delight in producing utilities and drivers that
make his ferrari notebook work better (I don't think he sleeps either).
Joerg Schilling does a lot of stuff with cdr - which I'm sure you know
about already. He's also produced an immense amount of code to power
SchilliX -- anybody who wants to is free to do likewise. Then there's
Murayama-san and his amazing collection of free ethernet drivers -
which I have been using myself for several years... the list goes on.
If you want to contribute what you develop back to the OpenSolaris
(and general OpenSource) community then that is great, fantastic even.
If you don't then that's your choice, but I seriously doubt that anybody
would really mind. These are just some examples of developers not being
driven by Sun's product and customer requirements, but producing what
they need and want, and making it available.
I think Sun is moving towards a non-scalable model of centralized
control. Instead you should look at something like a "Development
Tree" and "Official Solaris Tree". The Dev tree should be all owned
and driven by community - People will put in whatever interests them
and whatever they feel is the need of the hour. Sun can choose to
pick up stuff from there to go into the Official Solaris tree -
whatever their customers need.
Well gee, give us time! It is taking a while to get everything out
there -- there's the odd legal agreement which needs tracking down
and beating into shape, and then there's the rearranging of several
teams which are quite large so that they get into the spirit of Open
Solaris more than they have been. Believe me, I correspond with and
observe quite a few teams, and people are champing at the bit to get
their code and our code out there and under CDDL. Check back in a
few months and see how far we've come. Remember, linux took a while
to take off too, even without having to involve any lawyers.
On scalability, I think you're wrong, but it's way too early to make
any definitive statement about it. If it turns out that you are correct
then I know (based on my experience with this company and its culture
over nearly 6 years) that such a model would get fixed, and faster
than you'd think possible.
Otherwise this whole thing will be very limited - to Sun's ideas of
what is correct and wrong and to Sun's limited customer interests.
Sun's customers' interests are in fact quite broad :) I know this,
having worked in a support role for some time and seen a whole planet-
load of variety in what people do with Sun hardware and software.
This considering we don't want 10 forks of OpenSolaris
why not? There are $bazillions of what I would call forks of linux,
or indeed, Unix(tm) in general -- I'm in favour of having as many
forks of OpenSolaris as necessary for people/businesses/organisations
find a need for. I want to see OpenSolaris as an embedded OS option.
I want to see OpenSolaris extended in a similar fashion to SGI's
big systems. I want to see what people reckon they can do with Open
Solaris -- because people are smart and inventive and neat stuff
like that gives me goosebumps.
> - if that
isn't the case nothing is lost - people will still fork and do
whatever interests them.
Exactly, and that is one of the goals of opening up Solaris -- we
know we've got a tonne of amazing stuff inside Solaris, and (imho
at least) it's way past time we got to sharing it with everybody
else. I'm looking forward to the day that the code I work on gets
put under CDDL.
Finally, thankyou for taking the time to provide constructive feedback
on Sun's processes -- we need this. We're trying to make (Open)Solaris
much more than the sum of its parts, and those infrastructure things
that you don't like quite possibly do need tuning. We're not out to
make things worse, we're out to make things better. You're part of
that - and so is RMS - because it's not about the end point, it's
about the journey and the conversations that we have along the way.
Best regards,
James C. McPherson
--
Pacrim PTS Engineer 828 Pacific Highway
Gordon NSW
Sun Microsystems Australia 2072
_______________________________________________
opensolaris-discuss mailing list
[email protected]
