You present a well-organized commentary; however, I must amplify, and
thus take exception, to some of your statements.
First: Linux and Torvalds. Some might compare Torvalds to Bill Joy who
left a Berkeley PhD program for work in the private sector; Joy had a
sound background in what was "known" at that epoch. By comparison, I
suggest one consider the Tanenbaum–Torvalds debate
(see
https://urldefense.proofpoint.com/v2/url?u=https-3A__en.wikipedia.org_wiki_Tanenbaum-25E2-2580-2593Torvalds-5Fdebate&d=DwIDaQ&c=gRgGjJ3BkIsb5y6s49QqsA&r=gd8BzeSQcySVxr0gDWSEbN-P-pgDXkdyCtaMqdCgPPdW1cyL5RIpaIYrCn8C5x2A&m=p5LpkUSrDNa-AR53evz49_bezk928Gx00qoMLYEf4ys&s=2Cu2yWlCn1CePb2Zo769L4rY45NC0zplZHSqlCXvP1c&e=
for an overview).
Why did monolithic kernel Linux, based primarily upon the
non-production-environment OS Minix from Tanenbaum used as an
implemented example for teaching OS at the undergraduate level, achieve
sector dominance over micro-kernel BSD-derivatives? History, ease of
deployment (BSD typically was built from source even for end-users,
whereas Linux was "executable package deployed" as with Microsoft, the
prevalent desktop environment vendor. Linux picked up many, many
end-user applications, whereas BSD was much more sparse. Although both
are "POSIX", without various adaptation layers (not originally deployed
or even properly available), BSD cannot run a generic Linux binary
executable.
The example of a "small" regional USA government supporting a distro
does not address the "amateur" status -- there are paid persons who have
professional-status appointments but who are not professionals in the
academic/research/engineering proper sense. One may observe this in the
present USA Executive Branch (presumably changing under the current USA
President-elect); political persuasions aside, one may compare Dr. Atlas
to Dr. Fauci.
Your comment upon "amateur" status of various persons who have made
major research/engineering contributions is not my meaning of amateur.
Oliver Heaviside ( https://urldefense.proofpoint.com/v2/url?u=https-3A__en.wikipedia.org_wiki_Oliver-5FHeaviside&d=DwIDaQ&c=gRgGjJ3BkIsb5y6s49QqsA&r=gd8BzeSQcySVxr0gDWSEbN-P-pgDXkdyCtaMqdCgPPdW1cyL5RIpaIYrCn8C5x2A&m=p5LpkUSrDNa-AR53evz49_bezk928Gx00qoMLYEf4ys&s=y1LzX5fWUYJGJYtRf68p5Uf-S9dlyQH1FzulLHH4NQA&e= ) did
not have an undergraduate diploma, let alone any formal graduate school
education. Thus, in some "academic" sense, he was an "amateur" -- but
in reality, he was a consummate professional who made significant
advances in both the implementation and underlying formalism (including
"new" mathematics) of the physics (as well as the engineering and
technology) of his epoch. It is the understanding, knowledge, skills,
and dedication that make a "professional", not necessarily "formal"
education and diplomata; self-education will suffice (although often
deny that person the opportunity). Thus, in my opinion, neither
Torvalds (nor Gates) is a Heaviside. The "amateurs" you mention are
much closer to a Heaviside.
As for the other comments you make, we can pursue these mostly off-list
if you prefer. I do note that some Rocky EL personnel you envision to
be "paid" developers. Full time? "Gig"? From where do you envision
the pay to come? With proper benefits (not required in those
nation-states that have social services and benefits for all)?
Take care. Stay safe.
Yasha Karant
On 12/17/20 8:04 AM, Lamar Owen wrote:
On 12/16/20 9:55 PM, Yasha Karant wrote:
... The question I raised still needs to be addressed: will Rocky EL
be done by paid professionals (as with SL or Springdale Princeton EL)
or will it be done by volunteers, some (many) of whom are "amateurs"?
I am very concerned about the use in a production professional
environment of an "amateur" port of RHEL. ...
Conflating "amateur" with a lack of quality and "professional" with high
quality and guaranteed support is provably fallacious.
One of the very first RHEL rebuilds, White Box Enterprise Linux, was, to
use your notation, a "professional" production, sponsored by and for the
Beauregard Parish Public Library in DeRidder, Louisiana (read "County"
where they write "Parish," it's a Louisiana thing); see
https://urldefense.proofpoint.com/v2/url?u=https-3A__distrowatch.com_-3Fnewsid-3D01205&d=DwIFAw&c=gRgGjJ3BkIsb5y6s49QqsA&r=gd8BzeSQcySVxr0gDWSEbN-P-pgDXkdyCtaMqdCgPPdW1cyL5RIpaIYrCn8C5x2A&m=JTBeF2QPN2-NB4l7sB0VdZhNuE_mxophQaMcRPYwn5E&s=se-D6Q6pwAPkByDwIbTumyo9JAE46Eo5L8V6yTTzYvY&e=
But being "professional" didn't guarantee success; the last release was
in 2007. The "amateur" CentOS ended up with far better support with
mostly volunteers. I have liked and respected the Scientific Linux
developers and their attitude for quite some time, but it honestly
wasn't a surprise to me when it was announced that there would be no
SL8. The SL community seems to expect long-term support for any
arbitrary point release; that is really unsustainable with a small staff
and budget.
"Amateurs" can afford to dedicate more time in some cases than
"professionals;" in my own field at $dayjob the whole science of radio
astronomy owes its very existence to a talented and persistent amateur
by the name of Grote Reber. Sure, Jansky made the initial discovery
while on Bell Labs' payroll (as a "professional" he had to follow his
employer's money and go to the next project); Reber did the legwork and
got others interested, paving the way for "professional" radio astronomers.
In another major area of physics, thermodynamics, medical doctor Julius
von Mayer was overshadowed by James Joule; it didn't help that von Mayer
was a medical doctor, not a "professional" physicist. (a good overview
of that history:
https://urldefense.proofpoint.com/v2/url?u=https-3A__en.wikipedia.org_wiki_Mechanical-5Fequivalent-5Fof-5Fheat-23Priority&d=DwIFAw&c=gRgGjJ3BkIsb5y6s49QqsA&r=gd8BzeSQcySVxr0gDWSEbN-P-pgDXkdyCtaMqdCgPPdW1cyL5RIpaIYrCn8C5x2A&m=JTBeF2QPN2-NB4l7sB0VdZhNuE_mxophQaMcRPYwn5E&s=p0ZIGrcPxwlbndK4YUIC_ynHLup-BPnuyhqss6Ez9pY&e=
).
In computer science (using the non-ACM generalized definition of that
term), well, all I need to say is "Linus Torvalds." The very kernel you
run was an "amateur" creation, and for a number of years had no
"professional" support. Likewise, the Debian distribution was started
by "amateurs" and still has many "amateur" contributors; Ubuntu, a
supposedly "professionally"-supported distribution bases its work on the
"amateur" Debian; a chain is no stronger than its weakest link, and if
any part of even a "professional" distribution is supported by
"amateurs" ... "professional" Linux distribution support is a house of
cards built on an "amateur" foundation. It reminds me of the reasoning
in Ken Thompson's Turing Award acceptance lecture "Reflections on
Trusting Trust" (
https://urldefense.proofpoint.com/v2/url?u=https-3A__www.cs.cmu.edu_-7Erdriley_487_papers_Thompson-5F1984-5FReflectionsonTrustingTrust.pdf&d=DwIFAw&c=gRgGjJ3BkIsb5y6s49QqsA&r=gd8BzeSQcySVxr0gDWSEbN-P-pgDXkdyCtaMqdCgPPdW1cyL5RIpaIYrCn8C5x2A&m=JTBeF2QPN2-NB4l7sB0VdZhNuE_mxophQaMcRPYwn5E&s=-rEo5cSVS2fhIGxF42uFd_CWmc6DGwZNL3uLrDtYeL4&e=
).
One problem with relying on "professional" staff is that the entity
paying that staff has direct oversight into how much time they spend on
those problems; the funding entity's goals and any particular end user's
goals may differ dramatically, and the goals of the funder will trump
the goals of the user. A second problem is that the same "professional"
staff can be hired away by another company. A third problem is that
"professionals" expect to be paid; where does the salary come from? The
fourth problem is since there is very likely to be fewer "professional"
staff supporting a revenue-negative project, each "professional" becomes
extremely important or maybe even indispensible, and the project might
have a hard time surviving a "bus incident" or even a major hurricane.
I've witnessed all four of these issues first-hand RIP Seth.
The problem with "amateurs" is that they can quite literally walk away
without it negatively impacting their livelihood, and they're going to
work on what interests them, whether it interests the end-user or not.
I've witnessed "amateurs" walk away, try to delete everything they ever
contributed, and get mad when folks wouldn't forget what had been said.
At least with "amateurs" you can afford more of them, and have backups
for when people do leave.
As far as Rocky Linux is concerned, there is a middle ground where you
might have some paid developers and some volunteers; nothing wrong with
diversity here. I would expect that, just like the Linux kernel itself,
that we'll see a mixture of paid developers and volunteers for Rocky Linux.
