[CentOS] CentOS 8 Stream: The Good, The Bad, The Ugly

Wed Dec 9 23:51:12 UTC 2020
Joshua Kramer <joskra42.list at gmail.com>

Hello All-

After reading and digesting a ton of community chatter about the
recent CentOS announcement I've come to the conclusion that there's a
lot of good about this, but there are also a lot of concerns that are
being ignored.  And nobody so far has stared directly into the eyes of
the elephant in the room.  So here goes.

The Good: From a technical perspective- both in the sense of "getting
newer software" and "technical community being more involved in
bugfixes, etc"- having *a version* of CentOS called "AppStream" is
fantastic. The various RedHat and CentOS folks who have been extolling
these virtues in blog posts and twitter feeds are right-on.  But from
responses I've seen, it appears to me that they think that these
virtues are enough to completely gloss over the complete and utter
clusterfrackas they've caused.

The Bad: No point releases.  There is POSITIVELY NO* REASON that they
can't have AppSream and still do point releases.  Brand new stuff
would go into AppStream, at some point they do a point release of
RHEL, then follow the normal CentOS procedure to spin a CentOS build
of that point release.  This is already a tried and true process.  It
will cost RedHat all of what, low five digits (if that) in developer
salary to do this.  Heck I'm sure some volunteers would step up to use
the existing infrastructure if RedHat didn't want to spend any paid
developer time on this.

The Ugly: I denoted "NO* REASON" above because there actually *are*
reasons that we are not privy to.
https://twitter.com/JoshuaPKr/status/1336744681716244480  Since RedHat
is not being transparent with this, we are forced to speculate and
remain bewildered at why they would make a decision that is going to
cost them so much in the long run.  The most common (and most likely)
theory is that some MBA somewhere in middle management saw all of this
CentOS being used in production environments (and otherwise downloaded
for free), and had the idea that if CentOS had its head cut off people
would just buy RHEL subscriptions.

That may happen in a few cases, but for the most part, that is NOT
what is going to happen.  By handling the CentOS situation in this
way, RedHat has branded itself as a company that acts in bad faith. If
a company acts in bad faith towards a community where non-monetary
value is exchanged, WHY would you trust that company to hold up its
obligations for contracts that are actually paid?  People are going to
do whatever they can to get away from RedHat.  Debian, Ubuntu, SuSE
will all benefit from this.  Even in cases where non-profits and other
similar clients "contact RedHat about options because Stream won't
meet their needs"- why would such entities have ANY reason to trust
anything that RedHat says to them?

There have been hundreds of other messages that describe exactly what
RedHat loses in this deal so I won't go into that here.  But branding
oneself as a "bad faith actor" is usually a terrible way to try to
pick up a little bit of subscription revenue.  In the end it's going
to be a losing scenario.  This is an absolutely UNMITIGATED DISASTER
from a marketing and community goodwill standpoint.

It can, however, be mitigated if RedHat backtracks, admits their
mistake, and affirmatively commits to support future CentOS point
releases.  I'll be interested to see how this turns out.