How many Fedora users are there?

How many Fedora users are there?

Posted Oct 19, 2006 12:22 UTC (Thu) by JamesEM (guest, #41209)
In reply to: How many Fedora users are there? by bojan
Parent article: How many Fedora users are there?

>By providing verifiable information to Red Hat executives,
>Fedora has a much better chance of surviving as a project.

Any RedHat exec with sense should never make that trade, if the
Company gets into difficulties. It was clever to separate the "Brand
Values" of Fedora and RHEL, but that doesnt make Fedora as "disposable"
as the free staff that _created_ the critical mass for RedHat
to be able to exploit the mainstream market. Many "useless burdens"
were discarded at RedHat Linux EOL like an empty booster stage.
Maneovering for profit and success is rarely a victim free exercise.

Suddenly. many enterprise level support people found themselves
downgraded to "enthusiasts" relegated to Linux Desktop development,
and seeking credible enterprise alternatives that are more open
and less ethically mutable. At a disadvantage in their own market.

Should RHEL maintain entirely separate R&D support? I went ballistic
on the selinux list when RHEL rpms were offered up for (free!) testing.
The place for that is Fedora, otherwise a US taxpayer organization is
being used to underwrite commercial corporate R&D and provide unfair
market advantage. Thats a complete no-no, here in the EU.

Besides, all RHEL testers have a right to expect a financial return on
any such effort, surely? That *is* the business model Redhat chose for
themselves!

This grates with the guarantee of open access to a distro that once
benefitted from FREE R&D, and testing. Corporate execs at RedHat did
indeed lose the plot. Its easier to seek forgiveness than permission.
But they'll never regain my trust.

I dont like being made a victim. RedHat "mugged" a lot of long term
supporters and contributers with their changes. The personal ROI for
supporting Open Source products is a difficult enough niche. They
"captured" a substantial third party support market for themselves.

Fedora was the only return on that. Not good enough. ( for me ).
I've rejected the entire edifice, in total disgust.

A substantial investment in my time testing and integrating early selinux
as S/RPMs in RH Linux 9 went down the toilet at the EOL of that product.

Never again. Now: an additional learning curve to gain the same
skill levels with Debian/Ubuntu; a burden that compromises MY ability to
make a return on many years of OSS experience. Time and money. I no
longer have alternative sources of income to subsidize free R&D, and
for sure I'm not inclined to do that at no cost for any for-profit
corporate!

Corporate reality: Metrics ARE necessary.
So: with a little imagination, make it an open participation exercise.
You'll get better data! No sneaky stuff, no need for it.

Perceived as a tinfoil-hat type by some, yet I'm perfectly
happy to submit ( and use personally ) unique IDs for builds I maintain.
How hard can it be to "open source" _behaviour_ in this regard?
If the user has complete control over how those IDs get exposed from
their systems, its a non-issue. If you dont want to participate, then
dont provide the information.
( There should be a clear choice, and no penalties for choosing ).

Separate metrics and tracking. They are not the same! There need not be
privacy invasion; IP as a unique identifier is useless as a metric.
Forget that approach.

How hard is it for a given distro provider ( Corporate business _or_ org )
to openly have a simple form that combines various unique things from a
particular build - disk ID, MAC NIC address(es) into one (big) hash that
has a simple index number on their database? System builder gets the unique
index number, the company bean counters and marketers gain useful metrics.

The sysadmin ( by index ID only ) is invited to participate. Volunteer the
status of your build ( test, disposable ), or mainstream use. Even better.
Heck, why not provide a "group" ID, with some simple inventory and journal
fields for people to use? Thats could be of mutual benefit, if implemented
properly.

Hmmm. Maybe the community can support an OSS project like this as a common
standard for all distro providers? Just a thought!

---

(Log in to post comments)