Various people wrote:
> 1. I like X ...
> 2. blah ... Fedora ... blah
Personal preference is always a matter of choice. I never discredit the
choice of others, as it is personal preference. I support a lot of
distributions, especially Debian-based distributions like Xandros as
well as Debian itself, and SuSE is making more commercial in-roads thanx
to Novell. This is all fine, as people often argue that "choice is
good." But why does one need to continually expense Red Hat?
Because the one thing I have continually butt my head up against is
people stating things about Red Hat Linux and, now, Fedora Core that are
simply _not_true_, or full of _incorrect_assumptions_. This is not
"choice," it is "versus." In other words, it is "marketing" and
"unintentional FUD." And when this is portrayed in the IT media, I just
have to wonder how much of a disservice we, the Linux community, are
doing to ourselves?
I don't know how many times I've seen myself get "reamed by the
majority" on a list as a "Red Hat apologist" only to have someone
mention off-list that they work for a major integrator and that I'm 100%
correct. This has also led to some excellent opportunities for myself,
even if only short-lived work. Some people call me a jerk, and while I
don't deny that I am sometimes abbrassive in some of my responses, they
should take note that I don't "bash" distros or feel the need to make an
argument "versus" just to merely profess my preference. In fact, I
consider the whole "versus" comments to be the main problem of the
commercial software world. Let's reduce it to technologies, not brand
names and products. E.g., a "ports" distro versus a "packages" distro,
not Gentoo v. Red Hat, or FreeBSD v. Linux in general -- because there
is _major_overlap_ that assumptions don't show (and are, therefore,
incorrect).
Red Hat "assumptions" start as far back as Red Hat Linux 5.0 and the
GLibC 2 change, which was a _major_ necessity. They happened again with
Red Hat Linux 7 and GCC 2.96/3 adoption forcing strict ANSI C++
compliance that broke compatibility with GCC 2.8/2.91.66 (the latter aka
EGCS 1.1.2). The backport of NPTL to 2.4 gave yet more avenues for
people to complain, yet NPTL was being forced in 2.6.** But many don't
recognize that with Red Hat's "plunge" into GLibC2, GCC3 and NPTL --
they forced developers to support inevitable changes that were going to
happen in the near future. Red Hat is just the messenger with these
".0" revision changes for "early adopters," but is rarely thanked as
such (even when they _do_ maintain older versions _specifically_ so
people don't have to convert for years -- far longer than most other
distros I might add).
[ **NOTE: I document most of these "ABI Compability" reasons here:
http://www.vaporwarelabs.com/files/temp/RH-Distribution-FAQ-3.html#ss3.3
]
They continued through endless comments on the Red Hat(R) v. Fedora(TM)
trademark details only made matters worse to this day. Red Hat(R) tried
to use the "Official" prefix and that didn't work. And then it came ot
a head in 2002-2003, largely because Red Hat(R) had serious issues with
the USPTO thanx to Sun and other vendors who claimed "Red Hat(R)" was
public domain." Why? Sun and countless others didn't license but left
all the Red Hat(R) trademarks intact in their distributions, resulting
in the disputes. But at the same time, even when Red Hat explicitly
stated companies like Cheapbytes.COM could still call their CD's Red
Hat(R) Linux, the demonizations still wouldn't end.
The only thing Red Hat is guilty of is assuming people wouldn't abuse
their trademark in a 100% redistributable version. They were wrong.
But while people now demonizing Red Hat for this, they don't bother to
notice that no other, major commercial distributor allows unlicensed use
of their trademark either in off-shoots. You don't see Sun sporting a
SuSE trademark on their Java Desktop, and even then Sun had to license
the rights to it anyway (unlike when they used Red Hat(R) Linux prior).
As Red Hat constantly tries to state all over their Fedora site, "it's a
return to Red Hat's roots" -- i.e., here's a trademark and distribution
system that offers everything we always did with Red Hat Linux.
Michael Tiemann can say one thing in an interview and the IT media will
report another. Red Hat released a flavor of Red Hat Enterprise Linux
called "WS" and then even sold a shrink wrapped version explicitly
called "Red Hat Professional Workstation" and the media still said they
were only selling server software. So probably the biggest and direct
example of Red Hat trying to convey one more time "Hey! We still offer
a desktop solution!" was in the _intentional_mislabeling_ of 10 and
50-license packs of RHEL WS as "Red Hat Desktop." And yet the IT media
still say Red Hat has exited the desktop.
Now people ask me to point to Red Hat's site to "prove" my statements
are "true." They use IT media articles to prove theirs are "true." The
IT media has been predicting Red Hat would "close up" their Linux for a
longest time and it hasn't happened. In fact, when Sistina decided to
"close up" their LVM and GFS technologies, Red Hat bought them out, and
re-released them GPL. And before that, when Mission Critical Linux
started to "close up" various NFS fail-over technologies, Red Hat bought
them out. And going even before that, when Cygnus considered "closing
up" various GPL offerings, Red Hat decided to finally buy them out too
(although it was for other reasons too, Red Hat had been considering it
for a long time). When has the media _ever_ shown these moves in a
positive light to Red Hat?
No, it was in fact the opposite. Many people predicted that Cygnus
developers would leave en-masse as Red Hat would turn their stuff from
GPL into proprietary. Didn't happen (actually, the opposite happened,
Red Hat tried to get into the embedded market with too much GPL focus).
In fact, to this day, most commercial companies love Red Hat for their
GPL-focus and community efforts, but think they take it too far from a
commercial support standpoint. Probably the biggest coup to date is Red
Hat's buying out the Netscape Directory Server from AOL for $20M, which
will go GPL no later than April 30, 2005. Netscape Directory Server is
at the foundation of Sun One's directory implementation, and I'm sure
Red Hat's move to GPL it for _all_ to use is a final statement that
"hey, OpenLDAP ain't getting there." We all continue to benefit from
the proper usage of Red Hat's IPO capital to ensure the best software
stays GPL or becomes GPL.
Which brings me to my final point. The Red Hat Fedora page has changed
and the "About" and "Objectives" pages now match _everything_ that was
Red Hat(R) Linux before. But people can't put it together. So, the top
5 things you _won't_ here Red Hat representatives say ... for obvious
reasons of marketing ...
5. We use the exact same release model for Fedora(TM) Core as Red
Hat(R) Linux prior, but we changed the terms 1 for 1 to something else.
In fact, most of our developers still call it "Rawhide" when they know
they are contractually obligated to call it "Development" now. In fact,
we merely made previously non-public resources, like the Red Hat(R)
Linux lists, now renamed Fedora lists, open to the public for
subscription by anyone.
4. We provide the systems, network infrastructure and personnel for the
entire, official Fedora project, and even if this currently being moved
to Duke University under an independent Steering Committee, this will
not change. It is also the #1 legal reason why we have strict "free
software" guidelines in Fedora, because it's a liability issue for us, a
US-based corporation, regardless. The Red Hat(R) Linux development
model was always open, and now we have formalized the method this is
done with the Fedora(TM) Core "Steering Committee."
3. Every single package in Red Hat Enterprise Linux, _every_ single
package, has a _direct_ Fedora version, right down to the package name
and number. Packages that are marked "EL" typically just have
unsupported locales removed**. The "EL" kernel may have some hardware
additions under NDA, but builds perfectly on Fedora**. As a result,
with exception of the kernel, you can create a 100% exact instance of
Red Hat Enterprise Linux with Fedora Core, by maintaining the exact
package versions.
[**NOTE: You can find these references in various mailing list archives
at Red Hat. They go into detail on how most "EL" packages differ.
Otherwise, the packages of Fedora Core and RHEL are the _exact_ same. ]
2. Following onto #3, this is why we release Fedora(TM), under new
trademark guidelines. Now anyone can take our best stuff, and build
their own distro -- right down to _exact_ Red Hat Enterprise Linux
packages. This is also why the "White Box" and other "rebuilds" of RHEL
from source are self-defeating. Because they have to now remove the
trademarks in the packages, whereas they are a non-issue with
Fedora(TM). That's the whole reason for it! RHEL exists for a reason,
which brings us to #1 ...
1. SuSE was first with the Enterprise Linux idea, and our customers
told us Red Hat Linux 6.2"E" was the wrong strategy. Do you really
think we are going to admit this? Heck, it's right in SuSE's marketing
literature! They were first, okay, we admit, our own clients liked it
better! So we had to change. After 2 years, we finally evolved Red
Hat(R) Linux into Fedora(TM) Core, to address this new, split strategy.
Regarding #2, I have now been at 2 _major_ Fortune 100 companies over
the past year. They are doing _exactly_ this, maintaining an internal
Fedora repository that matches their RHEL deployments, as well as
planning for future RHEL releases (e.g., Fedora Core 2 before RHEL 4
came out). Why? To tweak support for Oracle and other software on
potential "updates" to RHEL before they are released.
-- Bryan J. Smith b.j.smith@ieee.org -------------------------------------------------------------------- Subtotal Cost of Ownership (SCO) for Windows being less than Linux Total Cost of Ownership (TCO) assumes experts for the former, costly retraining for the latter, omitted "software assurance" costs in compatible desktop OS/apps for the former, no free/legacy reuse for latter, and no basic security, patch or downtime comparison at all.----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 17:32:02 EDT