On 13/11/2007, Paul Jakma <paul at clubi.ie> wrote:
> On Tue, 13 Nov 2007, James Carlson wrote:
>
> > What we come back to is that a common and minimal "reference
> > distribution," at least as you're describing, isn't what the rest
> > of the proponents (notably Sun's marketing) want to have here.
> > Instead, they want a single "known good" distribution that can be
> > proposed for all first-time users.  Ignoring that desire will, I
> > think, set us up for future conflicts of exactly this nature.
>
> It seems the "good for reference distro" and "good for marketing
> distro" are two seperate things.
>
> - One needs to be a reasonably complete distribution, for a wide
>       audience.
>
> - The other needs to be confined to the intersection of all^Wmost
>       OpenSolaris distributions.
>
> The smallest, useful source subset of OpenSolaris available today
> would be OS/Net, no? The smallest, useful binary subset would be the
> reduced networking cluster of ON.
>
> Is there a good reason to not use those as the initial reference, and
> incrementally approve on them?

Having a reference distribution is one thing; having that reference
distribution be used to ensure compatibility is another.

With a reference distribution that is only used as a "minimal
bootstrap reference", NOT for compatibility, it is probably fine to
omit things such as a desktop environment, non-free components, etc.

With a reference distribution that is used as a guide for
compatibility, choosing a desktop, default libraries, etc. are all
extremely important for ISVs, etc.

As an independent developer, I want to be able to distribute binaries
that I *know* will work because all of the dependencies will already
exist on the system in a default installation. Anything that requires
the user to download additional pieces before my app will run is
unacceptable to me.

That's the mess I have right now with GNU/Linux when I distribute
binaries and I'll gladly turn my handy dandy flamethrower on anyone
who suggests it should be the same way :)

-- 
Shawn Walker, Software and Systems Analyst
http://binarycrusader.blogspot.com/

"We don't have enough parallel universes to allow all uses of all
junction types--in the absence of quantum computing the combinatorics
are not in our favor..." --Larry Wall

Reply via email to