Subject: Ports / pkgsrc unification ideas (fwd)
To: None <port-amiga@netbsd.org>
From: John Klos <john@sixgirls.org>
List: port-amiga
Date: 09/03/2000 18:41:10
Daemonnews had an interesting editorial about unifying Net, Free, and
OpenBSD's ports and packages:
http://www.daemonnews.org/200009/editorial.html

I was interested, so I sent the following letter to Chris. Is anyone else
interested in creating such a project? If so, we'd have the m68k NetBSD
section all wrapped up.

If I get enough responses, I may start a separate mailing list. Please let
me know!

John Klos

----------------

Hello, Chris,

I read your ideas about unifying Net, Open, and FreeBSD ports & packages,
and I'd like to suggest an idea that may be worthy of inclusion in such a
project.

Looking at any of a number of packages, one sees that most platforms do
not have an up-to-date binary available, or the port / package has not
been tested with a particular platform.

I would like to suggest, generally, that some sort of "build farm" be
established, so that people who work on a port can "send" a port / package
to "the farm", where a fleet of volunteer machines across the Internet can
take ports and packages and try to build them. Results can be sent back
automatically to the "farm" server.

Sourceforge has a compile farm, but it is not an automatic one, nor does
it represent any significant representation of machines (they're mostly
x86). The idea is good, though; make different machines available to the
developers to speed up the process.

Personally, I am very comfortable with the security of my machine, and I
would not hesitate to create an account for developers who want to test
their builds on a 68060 NetBSD machine.

What do you think of this idea? Should this be a consideration in the
planning of the developer's side of a unified ports / packages system? Or
should it be a separate project?

I have a number of programmer friends, in addition to myself, who would be
willing to work on such a project. I am willing to donate any needed
resources on my machines as well (I have several BSD machines colocated;
one is a production server for hosting and development, and more are
coming). I'd also be willing to create a "farm" server, and start fleshing
out details of how such a thing would be done.

I hope to hear from you soon,
John Klos
-- 
To those accustomed to the precise, structured methods of conventional
system development, exploratory development techniques may seem messy,
inelegant, and unsatisfying.  But it's a question of congruence:
precision and flexibility may be just as disfunctional in novel,
uncertain situations as sloppiness and vacillation are in familiar,
well-defined ones.  Those who admire the massive, rigid bone structures
of dinosaurs should remember that jellyfish still enjoy their very
secure ecological niche.
                -- Beau Sheil, "Power Tools for Programmers"