3rd party repositories vs the floppy disk

Bryan J. Smith b.j.smith at ieee.org
Fri Jun 29 18:19:38 UTC 2007


On 6/20/07, Karsten Wade <kwade at redhat.com> wrote:
> When you have created a package repo that can support all those
> personality types, then you have a model for

Communism is an utopia.  People forget that some of the best and
greatest communist utopias -- city-sized/wide -- were founded in the US.
Utopias work when everyone agrees and holds each other accountable.  You
don't need processes or rules or denials, because everyone just works
together.  For an engineer like myself, communism is the ultimate ideal
and efficiency -- which is why, not surprisingly -- the Soviet Communist
Party was dominated by engineers.

And it utterly _fails_ when you have an organization large enough were
personal accountability is not feasible, people _differ_ on goals.  And
that eventually results in another structure, be it of merit
(meritocratic-republic) or majority (democratic-republic) or otherwise.
And that's where others yet just won't deal with any "process" they
consider is "wrong."

How you address that, there's _no_ "silver bullet."  In the US, we
engineers studied microeconomics and its microcosm of technical
feasibility, market aspects and various limits -- despite it's gross
inefficiencies and sheer "unfairness" in comparison -- because we
believe capitalism, and the efforts people put towards its incentives,
is one way we know that works for us.  Not that it's perfect.  Not that
it doesn't stomp on people.  And Americans engineers will be the first
to tell you that "freedom" isn't about "fair."

I've had this argument with people in the CentOS project, various, major
maintainers like DAG and others, who constantly complain about the
"processes" and "inconsideration" that the Fedora Project imposes to be
a contributor.  But guess what?  Debian has its "rules."  Many other
projects have their "rules" as well.  And yes, it's driven somewhat by
an agenda of -- ultimately -- Red Hat, but that doesn't mean it's
necessarily a "bad idea."  Especially when it comes to not merely just
package-level testing, but more importantly, integration testing and
regression avoidance.

Those right there are very, very important characteristics of a properly
tuned engineering lifecycle.  While it would be great if we could all
work directly together like a single, engineering team -- and individual
software projects and packages included by and for Fedora can do that --
it still doesn't solve the larger issues of integration testing and
regression avoidance.

So what is the "utopia"?  I don't know, but I can tell you one thing for
certain.  When people _differ_ on what that is, it's a great sign that
it's _not_ possible at all.  So that's where the "rules" -- which we
engineers call things such as "requirements" and "specifications" and
"reviews" -- that matter.  And yes, market considerations -- oh those
inconsiderate, "evil enterprise" thoughts that Red Hat sometimes
"pushes" on the "community" comes out as Red Hat employees work on
Fedora, just like they do the kernel, GCC, GLibC, etc... -- are there,
especially when people don't feel they don't represent "the will of the
[Linux] people."

But Fedora, like any other "community" project survives on its
contributors.  And those contributors have built a set of processes and
requirements and foci base on their time, energy and efforts.  If people
who want to contribute, but claim they do not, because they claim
"Fedora," or worse yet, "Red Hat doesn't care" -- then they are free to
fork and do what they wish.  And we have continually seen what happens
in the majority of cases (although not all), those efforts die off.  And
the few that remain do remain because those contributors can work within
a set of rules and realizations that there's a lot of effort required.

"Utopia"?  That term ranks up there with my other favorite, non-tangible
term that people like to use as if it's a finite, discrete product or
service -- "Renewable."

On Thu, 2007-06-28 at 07:32 -0800, Jeff Spaleta wrote:
> I'm perfectly fine with trailblazer and pioneer repos, full of
> alternative builds or new things, that are intrinsicly more difficult
> to deal with. I'm not so fine with a set of tools aimed at settlers
> that encourages them to play in the more dangerous spaces without a
> liability waiver analogy.
> -jef"there is a reason that only 1 in 4 families who claimed land
> during the landrush of the mid-western united states were able to
> survive more than one winter."spaleta

Those who contribute and make Fedora are the ones who set the
"requirements" and "reviews" to be in Fedora.  They don't set those
details lightly, and they don't set them to "control" the process
without the utmost consideration for all.  Everyone is "free" to
contribute, as long as they respect those processes.  Those who don't
are free to consider other, alternative mechanisms and approaches
outside of those processes.

Some do, and they do a great job.  I'll always appreciate those who have
their own ways of making things work.

But those who complain about the "control" on Fedora or -- more
directly, Red Hat -- often use the same "it's not fair" I hear of so
many processes and systems.  It's _never_ been about "fair," it's only
been about consideration for those who came before them, what they
thought worked, and how they worked together -- "freely" not "fairly."

Most people agree what "free" is.  Few people agree what "fair" is.


-- 
Bryan J. Smith         Professional, Technical Annoyance
mailto:b.j.smith at ieee.org   http://thebs413.blogspot.com
--------------------------------------------------------
        Fission Power:  An Inconvenient Solution




More information about the marketing mailing list