Joerg Schilling wrote:
> "Garrett D'Amore" <gdamore at opensolaris.org> wrote:
>
>
>> Influence != control.
>>
>
> My basic proposal for a new ARC is that people who did already work on the
> same
> topic or something _very_ similar need to get more weight in their decisions
> that other people who may not even know the background in depth.
>
> The problems I saw have been a result of the fact that unrelated people have
> control while people with related skills are ignored.
>
No. People were involved with the decisions. Smart people listened to
what the folks who "know" were saying. Then made a choice based on
their understanding. In the case in point, you *had* (and still *have*)
the ability to attend the ARC meetings and have a say. You don't get a
vote, but you can try to influence those that do. And you can by
diligence and participation work towards becoming a full member who
*does* have a vote (this is something *I* am doing now -- working
towards full PSARC membership -- 'cause guess what -- even though I'm a
staff engineer at Sun, even I don't get a vote!)
ARC can't force you to participate. If you choose not to, please don't
blame the ARC for not making the decisions you want them to.
>
>
>
>> The community has numerous opportunities to influence and participate.
>> And while Sun needs to continue to make this more open. But the
>> community has failed to capitalize on all the opportunities that are
>> already available (such as more participation on ARC by non-Sun employees.)
>>
>
> Sun has failed to take the code opportunities that are available from people
> outside Sun. There are a lot of code extensions an bug fixes available that
> are
> ignored.
>
Not every opportunity has been taken -- but that is not the same as
saying none have been taken. And more and more are being handled every day.
Again, I think you have a particular beef because /bin/tar != star, and
because we haven't taken everything that Joerg Schilling wrote and
shipped it as part of Solaris. If you've followed other cases, you know
that other contributions *have* been taken, based on what priorities are
and based on the availability and participation of 3rd parties (and also
Sun parties) in getting all the various hurdles and process-work done.
If you want your stuff integrated, there are steps you can take. Have
you integrated your stuff into a Nevada or SFW workspace? Have you made
sure that all your stuff can pass full SPARC and x86 builds without any
lint or cstyle warnings? Have you asked for public code reviews? Have
you documented a test plan? Have you done the work to ensure your stuff
is ready for G11N (globalization)? What about A11Y (accessibility)?
Have you posted up test results?
The other bits of code that have been integrated from other 3rd parties
(namely, ksh93, afe, mxfe, and sfe) all went through this process.
Sun employees can help figure some of these steps out, but to be quite
honest, it helps if there is someone with a particular itch that your
software scratches. Right now, I *suspect* that many folks have higher
priority itches than another replacement for /bin/tar, or even
/etc/rmt. If you want your particular itch to gain higher priority to
the Sun folks who can help you, it will help if you can get a following
of additional folks who are also asking for the same feature. Right
now, I think most folks would prefer I (as just one example) spend my
time improving platform drivers, working with 3rd parties like Masa-san
to integrate additional driver support, and working on x86
suspend/resume support rather than working on replacing /bin/tar.
Sun is not going to just open the flood gates and let any
"self-proclaimed expert" integrate whatever they want, without any
controls or process, or checks. Nor should they. I suspect a straw
poll of the community would find that the community *prefers* to have
the controls that Sun establishes in place to ensure that both the
*architecture* and the *implementation* are of the highest quality that
we can reasonably assure.
Folks that prefer a more cavalier approach to software engineering and
integration may find other FOSS projects more to their liking.
- Garrett
