Joerg:

>> You've been told how to prevent new conflicts with your tools arising:
>> run one or more ARC cases for integrating your tools into OpenSolaris,
> 
> As long as it has not been verified that arc cases help, I will definitely not
> do this!
> 
> You are responsible to allow approved arc cases to be integrated before 
> arc cases can be taken for serious. As for today arc cases are a big tool
> to spend time on but they do not result in facts.

An interesting bit of history.  For a long time the Sun JDS GNOME
project team felt similar to you, that ARC is a waste of time,
doesn't produce results, etc.  A part of this bad attitude stemmed
from the reasonable fact that the GNOME stack is huge with many
complicated interfaces, making it painful to document as ARC
requires.

However, the JDS GNOME team bit the bullet, and I think the GNOME
ARC cases are a nice example of how a free software project and ARC were
able to work together.  In these reviews, numerous issues were
identified and resolved that otherwise would have been overlooked.  I
can honestly say that having an experienced group of architects look
over our proposals greatly improved the quality of the GNOME desktop
(both in general and how it is shipped on Solaris/OpenSolaris).

My experience has been that ARC is very agreeable to work with people to
figure out new ways of doing things when needed.  It took us several
failed attempts, but working together we found a way to document
interface change in the GNOME desktop that is suitable to ARC and
which also isn't so burdensome to the resources of the JDS team.

I understand that figuring out how to make ARC work with external
projects is a new thing, and will likely be a bit complicated at first.
Since I was involved with the pain of ARC'ing GNOME, I can sympathize
with anyone's tendency to try and avoid such pain.  However, it
shouldn't be a mystery to anyone that ARC is the review body that
determines which interfaces are considered blessed.

If you haven't even bothered to propose your interfaces to ARC (or
process changes if you think the ARC process needs improvement), then
you can't really claim to have tried very hard to work with the
process.

Brian
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Reply via email to