14 Jul
2019
14 Jul
'19
2:52 p.m.
On Sun, Jul 14, 2019 at 12:13:35AM -0700, Ed Carp wrote:
On Sat, 13 Jul 2019, space aliens made Dave Horsfall
write:
+1
The term "Open Source" dates to 1998, so saying the movement dates
back to 1992 is at best historical revisionism. If what you mean is
the concept of distributed development, enabled by the internet, and
you don't want to count Linux (which started in 1991), I'd point you
at perl from the late 80's.
Larry Wall was extremely welcoming to enhancements from people that he
didn't know except for the fact that they sent patches that passed
technical muster. Even if that person was a random undergraduate
systems programmer at MIT...
Both Larry Wall and Linus Torvalds subscribed to the "release early,
release often" methodology --- which was especially important in the
days before distibuted source control systems. If you want to
encourage contributors, it's really important that they get positive
feedback very quickly. So feedback on proposed patches, and letting
people see their contributions show up in a new release is
super-important. And that means releases on a schedule measured in
days or weeks, and not months.
So if anything, I'd claim that 386BSD was a great, early example of an
open source anti-pattern. Releases every six months might be fine if
you're using a physical distribution medium, like CD-ROM's, but one of
the key aspects of the "Open Source movement" was the distributed
development methodologies that was enabled by the 'Net.
- Ted
P.S. There are plenty of other comp.sources.unix and
comp.sources.misc "open source" projects from the 1980's, but Perl is
one of the much larger, much more visible, and with a very large
contributor base, which makes it a very early project that looks like
what many people think of when they say "a successful Open Source
project" today.
386BSD was released on this day in 1992, when
William and Lynne Jolitz
started the Open Source movement; well, that's what my notes say, and
Not really. Bill and Lynne kept very tight control over releases - the word
"open" didn't really apply to 386BSD, and there were many Open Source
projects well under way before 386BSD wasy even conceived.
Under Linux, the process was a lot more "open", even democratic. One of the
reasons I abandoned 386BSD early on and started working on Linux was because
I (as well as many others) were very frustrated at the complete contol the
Jolitz's exnercised over 386BSD, and limited releases to one every six months
- much slower than was generally considered to be acceptable for the long
list of bugs and fixes in the pipeline.