Stephen Hahn wrote:
>   Well, this seems like an entree to make my economic point.  We  
> cannot
>   staff each OSS integration effort such that it could modify a
>   significant subset of commands (or the system) to meet policies  
> and/or
>   advised changes on initial integration; the freeware team is at best
>   able to choose to ship or omit components.

I will add that the authors of the open source software do not want
redistributors to make independent changes to their code aside from
the minimum necessary to compile the code on a new platform.  The
authors end up having to deal with the issues that are created when
misinformed changes are made to their code.

If there are bits of Solaris policy that would be enhanced by an
alternative implementation within an OSS project, then the change
requests can be sent upstream.  Changes based on sensible policies
(i.e., those that reduce complexity or improve performance) are
usually adopted even when they are platform-specific.

I think the ARC should understand that attempting to pick
and choose from open source alternatives while making changes to
those products is not a feasible course of action.  It is okay to do
what the case suggests and install the original OSS product in a
location that can be selectively included in paths.  It is also okay
to take an OSS product, rename the product and all of the interfaces,
and then place whatever constraints you want on that new product.
At worst, that will only waste developer time on the fork and Sun's
customers can download the real product from the source.
It is not okay to pretend to integrate someone else's product,
including their names, while not actually doing so.  That results
in costing the original developers' time and reputation.  Apache
would consider it a trade name violation.

I hear a lot about Solaris policies and how so-and-so would violate
some ancient criterion learned from the olden days (invariably, the
ones involving Solaris 1-2 transition history).  I don't hear much
about the goal of improving Solaris as a platform so that non-Sun
developers will consider actively supporting it again.  The benefits
of having built-in support for GNU coreutils should be weighed
against the cost of duplicating some features in the overall
operating system.  I think it is a clear win, even if it would
increase the cost of export checks or FIPS compliance (which
I highly doubt given the legal-centric nature of those processes).

In short, I agree with Stephen's assessment that, given the
potential technical conflicts have been addressed by the case,
the remaining decision is economic in nature: does Solaris want
to tradeoff the benefit of having a working coreutils versus
the cost of duplicate code?

Cheers,

Roy T. Fielding                            <http://roy.gbiv.com/>
OpenSolaris CAB

Reply via email to