[ale] Maddog’s take on recent Red Hat source distribution changes

[Steve Litt]

Jon "maddog" Hall said on Sun, 20 Aug 2023 17:20:26 -0400

>Hi Steve,
>
>You, like so many other people in this argument, appear to be assuming
>that everything in "Linux" is GPL V2.x.   Mostly it is only the kernel
>that is GPL 2.x.   Many other packages in the distribution are other
>types of licenses, produced by many other entities.   I have not
>looked at RHEL for a long time, but I am willing to bet there are
>packages that are solely written by Red Hat, copyrighted by Red Hat
>and licensed by Red Hat perhaps under the GPL, but perhaps under some
>other license.  There are other packages that are under some
>"permissive" license like BSD, and have no requirement to send along
>the source code to the end user or anyone else. Even if there were
>only packages in RHEL that GPL, the entire distribution, the set of
>bits required to install and qualify as RHEL could have a different
>license.   

I publish most of my stuff under the Expat license (
https://directory.fsf.org/wiki/License:Expat ), which resembles one of
the BSD licenses. I do this knowing full well that Red Hat or Microsoft
or Oracle or Amazon can take some or all of it proprietary. When I want
to prevent this, I go GPLv2, which I like for its simplicity.

>In the past the only thing that was required was the
>removal of the Trademark information.

Of course. This is the whole point of trademarks.

[snip]

>My point in writing the article was to bring back to peoples minds the
>original reasons for Free Software.   The fact that people would get
>binary code and not be able to do what they needed with it.  The fact
>that companies would release binary code and then either force you to
>have a maintenance contract with them to get the bug fixes or
>enhancements you need.   The fact that over time the companies would
>walk away from the release you were counting on and never patch it
>again.  Can you say "Windows XP", with close to 12,000,000 systems
>still running it?  

Another original reason: So you could grab somebody else's free
software, modify it as desired, and give it free to 1000 of your
closest friends and customers, with no compliance hassles.

[snip]

>As to systemd vs init files, I have no bones in that argument, just as
>I stay away from vim vs emacs discussions.   

This is the crux of the situation. It isn't, and never was, systemd vs
sysvinit (which I assume is what you meant by init files). Overlooked
were runit, s6, OpenRC, Busybox Init, and several others.

I dislike the daemon control in sysvinit so much that back in the mid
00's I was using sysvinit for PID1 and to launch a few daemons, and
Daemontools to supervise the rest of my daemons.

As far as vim vs emacs, that's no longer a valid comparison either,
with VSCode, Bluefish, and many other great editors.

[snip]

>Finally, at the bottom of my blog article, was the real meat of what I
>had to say.   In the aftermath of Red Hat's announcement four
>different groups came forward and said that they were going to create
>a clone of RHEL.   If these four groups were going to spend the time
>and money to each create a clone, that means we would have four more
>"RHEL"s.   I pointed out that this was, at best, three clones too
>many.   I also pointed out that what we really needed was a competitor
>to RHEL, a better RHEL, not just a clone.

The competitor wouldn't have the gigantic marketing power of the
IBM/Redhat conglomeration, so no matter how good it was, it would be
hard to compete. You're a historian, remember "you can't get fired for
buying IBM?" They were still mouthing that platitude in the mid 80's
when I busted into the kitchen table programming sphere.

If people would buy it, it wouldn't be hard (relatively speaking) to
create such a beast from the very simple Void Linux. Just pay a few
people to maintain Void packages for the softwares that are essential to
the new solution provider, and you could produce something pretty darn
good. Of course, not having the intentional complexity of Red Hat, you
wouldn't get as much consulting and training work.


>
>But that is hard work, and from my experience would take much more
>money than the 10 million dollars that SUSE was putting forth, even if
>they only managed to copy the RHEL release source code (all of it, GPL
>or not) and rebuild it.   It is the rest of what is needed that would
>cost a lot more to bring to the table.   It probably could be done.  I
>would encourage it to be done.

From a technical standpoint I see it as very doable, but nobody's going
to risk 10 million dollars on the line when it's likely they'll be
outmarketed by Red Hat.

>
>Sorry if I gave you any other idea of what I was saying.

No problem.

SteveT