After 4 years, I've gotta say it. I understand there are some legal issues or tardemark concerns showing the lineage of Red Hat Linux to Fedora Core. Plus there are also marketing aspects on Fedora v. Enterprise. But the general ignorance of those who have never followed the Red Hat / Fedora development line, or left after Red Hat Linux 9, constantly want proof and debate me.
We *NEED* Red Hat, or at least via Fedora, to put up a page - linked on the front page - that covers both the 2-2-2 month Rawhide/Development-Beta/Test-Release cycle, followed by the 6-6/6-6-6/6-6-6-6 month release cycler of community to enterprise release (every 2-4 community releases). This would include the full plot going back to Red Hat Linux 4.0, as well as the first Red Hat Linux 6.2E (retroactively called "Enterprise 1"), and how that *PROVEN* lifecycle of package, integration and post-release test has held up. Especially with the regular cycle of community
I'm tired of having virtually *NO* pages but my own to refer to on this, to the point I've been writing a book on Enterprise Configuration Management: Fedora/Red Hat Linux (of which I've had not time to complete in the last 18 months worth of 60+ hour/weeks of consultant, not including travel and other time). I used to maintain an unofficial FAQ on this, and I should probably re-blog an update of it.
Fedora Core is not more "Beta" than Red Hat Linux was before it. And until Red Hat clarifies that - Enterprise Linux marketing be damned - those who want to demonize it that way will continue, with nothing professionals like myself can point to - other than our own publications or blog articles on the matter.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
On Thu, 2007-06-21 at 13:46 +0000, Bryan J Smith wrote:
Fedora Core is not more "Beta" than Red Hat Linux was before it. And until Red Hat clarifies that - Enterprise Linux marketing be damned - those who want to demonize it that way will continue, with nothing professionals like myself can point to - other than our own publications or blog articles on the matter.
I agree. Fedora is a weird case where reality is better than perception.
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
* Stable desktop and servers * Three years of security updates * Branding program with hardware manufacturers
I think you could bring back a lot of the independent consultants who went to CentOS or elsewhere after the transition from Red Hat Linux. It could also quiet people who say cutting-edge = never-better-than-beta.
-- Chris Negus
On Thu, 21 Jun 2007, Chris Negus wrote:
I agree. Fedora is a weird case where reality is better than perception.
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
I think you could bring back a lot of the independent consultants who went to CentOS or elsewhere after the transition from Red Hat Linux. It could also quiet people who say cutting-edge = never-better-than-beta.
Red Hat only has so many resources to spare for long-term maintenance.
Why? Because, as it turns out, it's pretty hard to do well -- and for that exact reason, it's worth a lot of money to Red Hat to do it well. So we do it for RHEL, and we do it for big bucks.
There's nothing wrong with someone deciding to offer that kind of support. That was actually the point of Legacy -- and with all the tools available, a reformed Legacy group might be able to do that pretty well.
But here's the question: who would join that effort when they could just get the same benefits from CentOS? And the answer is, basically, no one. The emergence of CentOS, one could argue, was the final nail in the coffin of the Legacy project.
--g
People don't realize that trying to maintain updates on a release that is made every 6 months is unrealistic. As I accurately tracked, when Fedora Core 1 was released, Red Hat was supporting 6 simultaneous releases! The same thing happened mid-RHL 6.x release - all the way back to 4.2. Companies "standardized" on Red Hat Linux 7.1 and "didn't want to move" to RHL 7.2 - causing more release support issues.
That's not feasible. Furthermore, hardware and software vendors are unwilling to recertify every 6 month release. And they were the ones that pushed (as I understand it) for a Red Hat Linux 6.2"E" with SLAs. SuSE Linux Enterprise 7 followed that, then Red Hat with RHAS/RHEL 2.1 based on Red Hat Linux 7.2 - and the rest is history. Companies that want longer-term support have to share in the costs of that - especially for backporting changes to maintain ABI/API and avoid regressions or interface changes.
Trying to maintain a community release cycle of 6 months more than a year is not realistic. However, clarifying some things about Fedora's development - including making those "taboo" references to Red Hat Linux's history, would go a long way to finally *Smashing* the non-sense commentary and demonizations. Using a release numbering with a bit more "warnings" would help as well, but I'm sure I'm asking too much.
That's my point, nothing more. It would be redundant and a waste of Fedora resources. Other projects that have attempted such have found themselves over-whelmed. You need the complementary release approach. Especially for leveraging one as a profit model to support the other.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
-----Original Message----- From: Greg Dekoenigsberg gdk@redhat.com
Date: Thu, 21 Jun 2007 12:08:04 To:For discussions about marketing and expanding the Fedora user base fedora-marketing-list@redhat.com Subject: Re: Linux.org: The Feodra 7 Year Itch
On Thu, 21 Jun 2007, Chris Negus wrote:
I agree. Fedora is a weird case where reality is better than perception.
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
I think you could bring back a lot of the independent consultants who went to CentOS or elsewhere after the transition from Red Hat Linux. It could also quiet people who say cutting-edge = never-better-than-beta.
Red Hat only has so many resources to spare for long-term maintenance.
Why? Because, as it turns out, it's pretty hard to do well -- and for that exact reason, it's worth a lot of money to Red Hat to do it well. So we do it for RHEL, and we do it for big bucks.
There's nothing wrong with someone deciding to offer that kind of support. That was actually the point of Legacy -- and with all the tools available, a reformed Legacy group might be able to do that pretty well.
But here's the question: who would join that effort when they could just get the same benefits from CentOS? And the answer is, basically, no one. The emergence of CentOS, one could argue, was the final nail in the coffin of the Legacy project.
--g
On 6/21/07, Chris Negus cnegus@rucls.net wrote:
On Thu, 2007-06-21 at 13:46 +0000, Bryan J Smith wrote:
Fedora Core is not more "Beta" than Red Hat Linux was before it. And
until Red
Hat clarifies that - Enterprise Linux marketing be damned - those who
want to
demonize it that way will continue, with nothing professionals like
myself can
point to - other than our own publications or blog articles on the
matter.
I agree. Fedora is a weird case where reality is better than perception.
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
I think you could bring back a lot of the independent consultants who went to CentOS or elsewhere after the transition from Red Hat Linux. It could also quiet people who say cutting-edge = never-better-than-beta.
-- Chris Negus
-- Fedora-marketing-list mailing list Fedora-marketing-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-marketing-list
Wouldn't that be like using Fedora 4 with years of updates? Fedora 5 support is ending soon, which means that software updates are no longer being provided. Using a 3 year old Fedora + patches wouldn't make it secure, it would make it old.
Just my 2 cents.
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
We do, it's called Red Hat Enterprise Linux.
On 6/21/07, Jesse Keating jkeating@j2solutions.net wrote:
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
We do, it's called Red Hat Enterprise Linux.
Red Hat is not doubt the leader, But you do not mention the cost.
I'm going to mention some words of the writer very healthily: "Perhaps a lesson from the Debian project is in order: release no distribution before its time. Fedora 7 developers have released a distribution that's not ready for prime time."
Debian provides this type of structure in its distribution, maintaining a unstable version, must understood like "unstable safe" or a "stable risk", at the same time its to stay in bloody line of the technology adding efforts in maintaining a version completelly "stable", this does not means that Fedora cannot be used in production environment, (sometimes Fedora is considering as a experimental distributions) but I must recognize that it constitutes a risk, a sysadm know that a package must be consider stable "with time in the market production", its RHEL and CentOS and is here where maybe the branch stable Debian version wins. Fedora need a official stable release (as Debian stable or CentOS for RHEL) matters only for security updates version of choice for networks and servers for those for whom dependability matters more than the latest software, and on one hand the already customary itch out releases.
I believe that Red Hat I do not want to create a new competition of RHEL and this is the reason for which a branch stable of Fedora is impossible beyond of operative costs that imply to maintain it.
Wilmer Jaramillo M. wrote:
On 6/21/07, Jesse Keating jkeating@j2solutions.net wrote:
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
We do, it's called Red Hat Enterprise Linux.
Red Hat is not doubt the leader, But you do not mention the cost.
What cost? Rebuilds are available for gratis downloads. End users pay for support and services.
Fedora need a official stable
release (as Debian stable or CentOS for RHEL) matters only for security updates version of choice for networks and servers for those for whom dependability matters more than the latest software, and on one hand the already customary itch out releases.
Common misconception is that RHEL updates are only security fixes. This is far from true. There are some selected features backported into major update cycles. For example, RHEL 4.5 added limited para virtualization support.
I believe that Red Hat I do not want to create a new competition of RHEL and this is the reason for which a branch stable of Fedora is impossible beyond of operative costs that imply to maintain it.
If you mean "stable" as in relatively unchanging then Fedora isn't tied to that goal. The value of RHEL is precisely because there is some real costs in maintaining a commercial product with associated service level guarantees.
Rahul
I wish Debian users would stop and understand the reasons for the DCC release. Unchanging ABI/API and LSB run-time compliance for one.
Kudos to the Debian project in general, especially since they support more than one ABI/API "set" per release (I.e., different GCC, GLibC, kernel, etc...). Of course, those decisions come with their own costs - both in workload, release cycle and compatibility. But the negatives are also a major, driving reason behind the DCC releases.
I might be long removed from Debian now, and would not even attempt to dissect it. But one thing is the honest truth: the Fedora (and RHEL) and Debian (and DCC) release lifecycles are *NOT* even remotely comparable. In fact, as I detailed in my other post, Debian and many other distros have directly benefitted from RHL/Fedora by "waiting" on their "shake down."
Now if you really want to talk about "different," we can start looking a "ports" distros instead of "packages" ones. E.g., Gentoo It even applies far less and cannot be compared at all, no matter how much people try to fit Portage into the comparison.
When I want early/bleeding edge, source access, I'm at Gentoo - especially in initial development/adoption. When I want leading edge, but integration tested, software packages, I'm at Fedora - especially for core development/adoption. When I'm want proven, integration tested and regression and change avoidance, I'm at RHEL - especially for long-term release and support.
Debian offers you more flexibility, but I've also been hit by bad, unforseen interface changes and regressions in package updates years after initial release. It's hard, damn hard, to balance new features against proven, fixed ABI/API standards. That's why the split lead-edge v. backport packages distros - like RHL/Fedora v. RHEL "just work."
And when I need to get to pre-Fedora adopted (e.g., Rawhide-state or even earlier) source, then I have to hit the source, so I just go Gentoo. But only for select development systems or more embedded developments.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
-----Original Message----- From: Rahul Sundaram sundaram@fedoraproject.org
Date: Fri, 22 Jun 2007 04:19:11 To:For discussions about marketing and expanding the Fedora user base fedora-marketing-list@redhat.com Subject: Re: Linux.org: The Feodra 7 Year Itch
Wilmer Jaramillo M. wrote:
On 6/21/07, Jesse Keating jkeating@j2solutions.net wrote:
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
We do, it's called Red Hat Enterprise Linux.
Red Hat is not doubt the leader, But you do not mention the cost.
What cost? Rebuilds are available for gratis downloads. End users pay for support and services.
Fedora need a official stable
release (as Debian stable or CentOS for RHEL) matters only for security updates version of choice for networks and servers for those for whom dependability matters more than the latest software, and on one hand the already customary itch out releases.
Common misconception is that RHEL updates are only security fixes. This is far from true. There are some selected features backported into major update cycles. For example, RHEL 4.5 added limited para virtualization support.
I believe that Red Hat I do not want to create a new competition of RHEL and this is the reason for which a branch stable of Fedora is impossible beyond of operative costs that imply to maintain it.
If you mean "stable" as in relatively unchanging then Fedora isn't tied to that goal. The value of RHEL is precisely because there is some real costs in maintaining a commercial product with associated service level guarantees.
Rahul
Sorry, FUD.
1. The Debian (and DCC) and Fedora (and RHEL) release models are not comparable. Some of us have been Debian maintainers and we'll be the first to tell you that. ;)
2. Fedora is quite stable for production use, especially in an environment with AFS or NFS and other UNIX infrastructure. A lot of that has to do with Fedora's focus on enterprise features.
3. The release model for Fedora was _already_proven_ by no less than 15 Red Hat Linux releases. It remains unchanged now. The "release often" model is not an issue, it's a requirement of the lifecycle.
Why oh why do people have to compare Fedora to other distros? If you don't like the proven RHL/Fedora release model and lifecycle, at least stop to understand it. If you prefer Debian's lifecycle, by all means, go for it!
But let us not forget these realities ...
What distro put forth GLibC 2 (threaded libC)? What distro put forth GCC 3 (ANSI C++ compliance)? What distro put forth NPTL (various threading/compat)? What distro put forth SELinux (MAC/RBAC)?
People talk about Red Hat Linux, then Fedora Core, "breaking things." People are still talking about how Red Hat Linux 5 "broke everything." Same deal for 7, 9 and Fedora Core 2, without stopping to wonder why. No one steps back to realize the _massive_effort_ it took to adopt these radical changes. They just look at Debian or some other distro some year or two later and say - "oh, we waited until it was stable."
Um, no. Sorry.
You waited until another distro adopted something, including redeveloping many core components, and then all other developers/maintainers had to accommodate, and reaped the benefits of that after-the-fact. All while blaming it on Red Hat Linux, now Fedora, two fold - complaining about the change and incompatibility, then hypocritically saying you waited on the compatibility.
Sorry, putting something into Sid isn't the same as a major, refactored Red Hat Linux / Fedora release. It takes a crapload of effort to make massive, core changes, just to get to that release. And that still doesn't address the issues post-release of added software, which also has to be refactored. It's then, and only then, we get the necessary, concerted effort to "move forward."
Not by sitting back and waiting for someone like Red Hat to force the issue and projects to accomodate. Also remeber the percentage of contributions Red Hat developers make to various projects, to help them refactor and redevelop after major changes.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
-----Original Message----- From: "Wilmer Jaramillo M." wilmer@fedoraproject.org
Date: Thu, 21 Jun 2007 18:20:09 To:"For discussions about marketing and expanding the Fedora user base" fedora-marketing-list@redhat.com Subject: Re: Linux.org: The Feodra 7 Year Itch
On 6/21/07, Jesse Keating jkeating@j2solutions.net wrote:
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
- Stable desktop and servers
- Three years of security updates
- Branding program with hardware manufacturers
We do, it's called Red Hat Enterprise Linux.
Red Hat is not doubt the leader, But you do not mention the cost.
I'm going to mention some words of the writer very healthily: "Perhaps a lesson from the Debian project is in order: release no distribution before its time. Fedora 7 developers have released a distribution that's not ready for prime time."
Debian provides this type of structure in its distribution, maintaining a unstable version, must understood like "unstable safe" or a "stable risk", at the same time its to stay in bloody line of the technology adding efforts in maintaining a version completelly "stable", this does not means that Fedora cannot be used in production environment, (sometimes Fedora is considering as a experimental distributions) but I must recognize that it constitutes a risk, a sysadm know that a package must be consider stable "with time in the market production", its RHEL and CentOS and is here where maybe the branch stable Debian version wins. Fedora need a official stable release (as Debian stable or CentOS for RHEL) matters only for security updates version of choice for networks and servers for those for whom dependability matters more than the latest software, and on one hand the already customary itch out releases.
I believe that Red Hat I do not want to create a new competition of RHEL and this is the reason for which a branch stable of Fedora is impossible beyond of operative costs that imply to maintain it.
On Thu, 2007-06-21 at 12:47 -0400, Jesse Keating wrote:
On Thursday 21 June 2007 11:49:57 Chris Negus wrote:
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases?
We do, it's called Red Hat Enterprise Linux.
Yes, I should have expected that. Thanks Jesse ;-)
I guess I overstated my case. As someone who actively promotes Fedora, I sometimes feel frustrated that the quality the Fedora project achieves (with most releases) is overshadowed by some people's beliefs that you shouldn't really try to do anything with Fedora since it's unstable.
Maybe what I can do is write up a case study of what people actually have success doing with Fedora. Deployments in schools, homes or small businesses. In other words, Fedora project doesn't have to guarantee anything, but it can say what others have done. Perhaps someone could point me to some Fedora success stories to get me started.
Anyway, thanks for other suggestions on this thread to create and support my own distro from Fedora. That is not the least bit likely (although I did make a pretty cool Fedora 7 8GB install DVD image with one pungi command... thanks again to Jesse).
-- Chris Negus
On Friday 22 June 2007 10:53:29 Chris Negus wrote:
Maybe what I can do is write up a case study of what people actually have success doing with Fedora. Deployments in schools, homes or small businesses. In other words, Fedora project doesn't have to guarantee anything, but it can say what others have done. Perhaps someone could point me to some Fedora success stories to get me started.
This would be helpful indeed. Many places refresh every 12 months or quicker anyway so Fedora provides them a very useful target. A place on the wiki to start gathering such success stories would go a long way I bet.
I think it would be redundant and a side-track to offer such. RHEL offers SLAs and other options, options that go hand-on-hand with its approach. For those that don't need SLAs, CentOS becomes and option (with caveats).
Fedora continues to offer, as did Red Hat Linux prior, the leading edge in community development. Some releases offer less changes than others, with more proven integration and stability. Others offer more radical changes, which are required to move the codebase forward.
At the most, I wish Fedora would move back to the release numbering prior to Red Hat Linux 8. Bring back the dot (".") and the big-dot-oh (".0") warning that the wrench has been thrown. That's one thing I missed once Red Hat Linux 9 came out, and wasn't numbered 8.1. I don't care if NPTL killed some compatibility, the attitude has always been ...
X.0, we changed a lot X.1, only a few things changed (less likely RHEL fork) X.2, we really like what we have (more likely RHEL fork) X.3, oh, we could use another, the new stuff ain't ready yet
But other than that nomenclature/meta aspect, there is *0* I'd change. Maybe support the last revision longer than the others, but that's about it. In fact, this could be used to support .0 releases not as long, and support only things that are more stable, more proven longer.
But RHEL (and CentOS) don't need to be duplicated. Red Hat is 100% correct on how the segmented the lines, because each cater to the needs better of those they target.
Other than that, people didn't care about the trademarks issues then. So they aren't really likely to care about understanding anything now.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
-----Original Message----- From: Chris Negus cnegus@rucls.net
Date: Thu, 21 Jun 2007 10:49:57 To:fedora-marketing-list@redhat.com Subject: Re: Linux.org: The Feodra 7 Year Itch
On Thu, 2007-06-21 at 13:46 +0000, Bryan J Smith wrote:
Fedora Core is not more "Beta" than Red Hat Linux was before it. And until Red Hat clarifies that - Enterprise Linux marketing be damned - those who want to demonize it that way will continue, with nothing professionals like myself can point to - other than our own publications or blog articles on the matter.
I agree. Fedora is a weird case where reality is better than perception.
Has anyone on the Fedora project considered making one stable Fedora release every three or four releases? You could promote the release as having:
* Stable desktop and servers * Three years of security updates * Branding program with hardware manufacturers
I think you could bring back a lot of the independent consultants who went to CentOS or elsewhere after the transition from Red Hat Linux. It could also quiet people who say cutting-edge = never-better-than-beta.
-- Chris Negus
Bryan J Smith a écrit :
After 4 years, I've gotta say it. I understand there are some legal issues or tardemark concerns showing the lineage of Red Hat Linux to Fedora Core. Plus there are also marketing aspects on Fedora v. Enterprise. But the general ignorance of those who have never followed the Red Hat / Fedora development line, or left after Red Hat Linux 9, constantly want proof and debate me.
We *NEED* Red Hat, or at least via Fedora, to put up a page - linked on the front page - that covers both the 2-2-2 month Rawhide/Development-Beta/Test-Release cycle, followed by the 6-6/6-6-6/6-6-6-6 month release cycler of community to enterprise release (every 2-4 community releases). This would include the full plot going back to Red Hat Linux 4.0, as well as the first Red Hat Linux 6.2E (retroactively called "Enterprise 1"), and how that *PROVEN* lifecycle of package, integration and post-release test has held up. Especially with the regular cycle of community
I'm tired of having virtually *NO* pages but my own to refer to on this, to the point I've been writing a book on Enterprise Configuration Management: Fedora/Red Hat Linux (of which I've had not time to complete in the last 18 months worth of 60+ hour/weeks of consultant, not including travel and other time). I used to maintain an unofficial FAQ on this, and I should probably re-blog an update of it.
Fedora Core is not more "Beta" than Red Hat Linux was before it. And until Red Hat clarifies that - Enterprise Linux marketing be damned - those who want to demonize it that way will continue, with nothing professionals like myself can point to - other than our own publications or blog articles on the matter.
-- Bryan J Smith - mailto:b.j.smith@ieee.org http://thebs413.blogspot.com Sent via BlackBerry from T-Mobile
I have already beat you with early post =p
https://www.redhat.com/archives/fedora-marketing-list/2007-June/msg00186.htm...
marketing@lists.fedoraproject.org