"Considering the fact that every single package in rawhide is currently maintained by Red Hat employees, and we are diligently working every day to improve the Fedora Core release and fix as many bugs as possible, I'm not sure exactly how "Redhat just lost there ownership of the project" as you state above. While we have opened up the project to the community, the infrastructure is not yet in place to allow external developers to maintain or contribute packages, and so for all intents and purposes, Fedora Core 1, is very much a distribution developed by Red Hat not much differently than Red Hat Linux 9, or Red Hat Linux 8.0, 7.x, etc."
Currently they are all maintained by Redhat but since they plan on turning it over to the public more, that will be a harder task to watch what all these guys are doing to every package that comes out. Considering the fact that they are not even willing to support it any more for even the 30 day installation shows their loss of ownership. They keep close ties on the project so they have all testing on the packages for their commercial line. I don't believe that it will be as stable as the OLD REDHAT because they will be letting so many people touch it and too many people be involved that it will be more likely to have bugs. When Redhat was in full control they knew every change that happened. With so many cooks in the kitchen you will never know everything that was changed always leaving some mess around. They want to develop and let users use it for only the benefit of adding fully tested products to their enterprise edition. If they really wanted to merge with a group to make stable and more advanced version of Linux that everyone could enjoy, why not merge with a group more similar to Debian. They have a good and stable release but are lacking on the ease of use that someone like Redhat has. This would be a better and more positive move forward for the Linux community. They already have tons of developers that are working together on a common goal. So far the project might be the same as Redhat 9 but the future is what I am referring to. Like I said before TOO many cooks in the kitchen will lead to a big mess.
On Thu, 9 Oct 2003 David.Grudek@anixter.com wrote:
"Considering the fact that every single package in rawhide is currently maintained by Red Hat employees, and we are diligently working every day to improve the Fedora Core release and fix as many bugs as possible, I'm not sure exactly how "Redhat just lost there ownership of the project" as you state above. While we have opened up the project to the community, the infrastructure is not yet in place to allow external developers to maintain or contribute packages, and so for all intents and purposes, Fedora Core 1, is very much a distribution developed by Red Hat not much differently than Red Hat Linux 9, or Red Hat Linux 8.0, 7.x, etc."
Currently they are all maintained by Redhat but since they plan on turning it over to the public more, that will be a harder task to watch what all these guys are doing to every package that comes out.
That is not really true at all. Allowing external developers to get involved is not the same thing as "turning it over to the public". I've indicated that this is not going to be a freeforall where external developers do whatever they want to whatever package(s) they want and stuff them into the distribution. That is not in any way what the Fedora Core distribution will be comprised of. Red Hat _employees_ currently maintain all packages in the distribution, and will continue to do so for the forseeable future. Once the infrastructure is developed to allow externally maintained packages to be.. well, maintained, then certain packages in the core distribution may be offered up for grabs.
That doesn't in any way mean $JOE_RANDOM_DEVELOPER will be allowed to take random packages and maintain them with zero care towards quality control or established packaging and maintenance guideines.
It will very much be a meritocracy as described on the Fedora website, which means a developer or potential developer needs to show competance and skill in maintaining packages, and follow packaging guidelines and whatnot, being aware of project deadlines, and actually fixing bugs, etc. in the packages they've volunteered to maintain. If someone can't meet the requirements to maintain a package properly, they wont likely be maintaining packages in the project.
You're reading "Red Hat permits others to get involved" and misinterpreting it as "Red Hat hands over the project to others and lets them do whatever they want", and this is very much not what is happening at all.
Considering the fact that they are not even willing to support it any more for even the 30 day installation shows their loss of ownership.
You don't appear to have any idea what "ownership" means.
They keep close ties on the project so they have all testing on the packages for their commercial line. I don't believe that it will be as stable as the OLD REDHAT because they will be letting so many people touch it and too many people be involved that it will be more likely to have bugs.
Again, you mistakenly beleive that this will be a freeforall that any random people can jump in and take over packages and regardless of skill they'll be accepted with open arms and allowed to do whatever they like. This is very much not the case, as I've indicated above several times.
When Redhat was in full control they knew every change that happened.
Red Hat _is_ in full control. Once the infrastructure is put into place, people will have an opportunity to contribute to the project based on their skills, and based on meritocracy. The project is divided into several components as discussed on the website. Fedora Core is the base operating system, of which Red Hat employees currently maintain, and will continue to maintain for the forseeable future, at least the large core critical parts of the OS. If someone doesn't want to maintain a package any more, and an external developer shows competance, and is willing to take the package on, as long as the project believes the person is competant enough to do the job, they may be allowed to maintain the package. They'll be expected to follow project guidelines of course, and if they screw up or go off wildly doing random things with disregard to the project goals or something, they probably wont be maintaining things for very long.
It's not a free for all.
With so many cooks in the kitchen you will never know everything that was changed always leaving some mess around.
There wont be "so many cooks in the kitchen". Red Hat ultimately will decide what goes into the operating system, and others will be allowed to contribute to that. It's not a case of random developers deciding they're going to do this or that, and Red Hat having zero say, and Fedora effectively being hijacked by some crazy people on an insane package crapification rampage. This is a controlled project, which is very open to suggestion and discussion, but Red Hat, as claimed on the website will retain editorial control, and ultimately makes the final decisions about things. That is required in order to keep the project sane, and not have it wandering off in random directions.
They want to develop and let users use it for only the benefit of adding fully tested products to their enterprise edition.
I'm one of "they". Red Hat Enterprise Linux is already built on top of Red Hat Linux, and it has been all along. Red Hat Enterprise Linux will continue to be built on top of the Red Hat community operating system releases. Only now, the community OS release is now made a public project, in order to open up the internal processes of communication more widely to the community in order to improve the Fedora project by incorporating ideas from the community, and to allow others to get involved if they feel like doing so. Those who want to get involved, can get involved simply by making suggestions and coming up with ideas, and discussing them on the development lists, or they can do actual development and submit patches and enhancements if they like. If someone wants to maintain a package or packages, and can demonstrate compentancy for doing such, and for following project packaging and development guidelines, they may have the opportunity to do so. This is a possiblity both for the core OS (which was previously Red Hat Linux), and also for the addon repository(s) Fedora Extras, Fedora Alternatives, etc.
Our goal is to take the same concept that has made open source software work well, and extend that to the distribution as a whole. Looking at GNOME as an example. The GNOME project is maintained by GNOME developers, and it has a steering commitee which guides the development, etc. GNOME is developed by many many people, and it is a high quality open source project with lots of high quality applications. That project's code gets incorporated into Red Hat OS products, and most of it is developed by the community, by skilled developers.
The same concept can be applied to an entire operating system distribution, and that is what we are trying to do here. Red Hat currently fills the role of "steering commitee" above, and currently also as "developer community" since there are not any external developers currently maintaining packages in the distribution core yet due to the infrastructure not being in place yet. Once that infrastructure is in place, then as people show interest, and also skill to contribute to parts of development, and volunteer to do so, we hope to add their skills to the project and let the project grow.
And just like the GNOME project's efforts get included in the community based OS, as well as the enterprise based OS, the community's efforts on Fedora Core will be used in future enterprise products as well. Why wouldn't they? ;o)
The general idea here is that with proper controls and project leadership, and a direct set of project goals, a community of external developers so inclined to contribute to the project can do so, and can help shape the OS. The results of that will be the community Fedora Core release of the OS, and the community maintained Extras and Alternatives addon bits.
If they really wanted to merge with a group to make stable and more advanced version of Linux that everyone could enjoy, why not merge with a group more similar to Debian.
People who want to use or contribute to Debian and their developmental model are of course free to do so. The Debian project and it's development model however doesn't meet the goals of the Fedora Project any more than the Fedora Project meets the goals and development model of Debian.
They have a good and stable release but are lacking on the ease of use that someone like Redhat has. This would be a better and more positive move forward for the Linux community.
There are many ways to contribute in a positive and useful manner to the community. Red Hat has been doing so for years now, and with the new Fedora Project direction, hopes to increase this even more.
They already have tons of developers that are working together on a common goal. So far the project might be the same as Redhat 9 but the future is what I am referring to. Like I said before TOO many cooks in the kitchen will lead to a big mess.
Using your own example, the Debian project has numerous developers, and they also have project goals and guidelines, and rules and regulations, as well as project leaders. The Debian project seems to have done well for itself, and has proven that a successful Linux distribution can be developed and maintained by the community. Do they have too many cooks in their kitchen? Or are they an example rather that shows that a community of competant developers are able to contribute collectively to a common project with common goals and produce something very useful as a result?
There is no reason why the Red Hat Fedora Project can't open up development of the base operating system formerly known as "Red Hat Linux" and create a successful project using similar principles but different project goals, and provide our own set of guidelines and rules in which the project is ran - to be influenced of course by those interested in contributing as well.
The only cooks that are in the kitchen currently, are Red Hat employees in this case, and some of the fedora.us people whom we've joined forces with, so your analogy holds no water I'm afraid. You're criticizing a project without fully understanding it at all, using broken analogies that don't apply to how the project will be ran, nor the goals of the project.
Feel free to be skeptical all you like, but because you don't understand something doesn't mean it can't or wont be successful on it's own merits. I suggest you read every single web page on the http://fedora.redhat.com website, and then ask specific questions on the list to clarify any misunderstandings, misconceptions, or questions you may have.
This is about positive change, harnessing the open source methodologies that brought us things like the Linux kernel, XFree86, GNOME, and many other highly successful projects, and using some of these methodologies in the creation of an entire operating system. However, we're not cloning existing efforts out there such as Debian, Gentoo, or other projects. Those projects and others have very definitive goals of their own, and those goals differ from our projects goals significantly in some ways, while we also share some goals as well.
On Thu, 2003-10-09 at 09:27, David.Grudek@anixter.com wrote:
am referring to. Like I said before TOO many cooks in the kitchen will lead to a big mess.
No, because Fedora gives us a bigger kitchen! It's called "scaling".
On Thu, Oct 09, 2003 at 05:13:03PM -0600, Bill Anderson wrote:
On Thu, 2003-10-09 at 09:27, David.Grudek@anixter.com wrote:
am referring to. Like I said before TOO many cooks in the kitchen will lead to a big mess.
No, because Fedora gives us a bigger kitchen! It's called "scaling".
Besides, we already have something with a lot of cooks involved. It's called "all the packages that make up Fedora". Open source projects - and the Linux kernel is a prime example - can benefit from having "too many cooks" - it can be somewhat chaotic but in the long run it produces better software.
What Fedora should bring to the table, just as RHL did before it, is _more_ QA, _more_ dependability, than you would get from just installing the latest versions of each package from the upstream.
(Although, admittedly, the Fedora model is to try and make changes upstream only as much as posssible, for most packages. I suspect that things like the kernel would be an exception to that guideline, for the simple reason that enterprise customers want to see things backported from development to stable kernel series, and the enterprise kernel will need real-world testing which Fedora can provide. Not being a RH employee I couldn't say for sure, of course.)
-- Robin
On Fri, 10 Oct 2003, Robin Green wrote:
(Although, admittedly, the Fedora model is to try and make changes upstream only as much as posssible, for most packages. I suspect that
I don't really see how that is any different than we do now though. For example, I'm not going to not fix bugs in XFree86 because XFree86.org hasn't checked a patch into CVS. It could take 8-12 months for that to happen. Right now my XFree86 rpms have over 120 patches in them, and future builds will have more and more. If it didn't then what would be my job for example?
We don't really control upstream projects much, all we can do is make sure to send our fixes and whatnot upstream. But we certainly should never wait until upstream projects apply our fixes and release new versions of their software in order to ship it. That is just silly and needless lag. The distro would never be stable and/or would never ship.
Patches are and always will be needed to fix bugs.
Mike A. Harris wrote:
On Fri, 10 Oct 2003, Robin Green wrote:
(Although, admittedly, the Fedora model is to try and make changes upstream only as much as posssible, for most packages. I suspect that
I don't really see how that is any different than we do now though. For example, I'm not going to not fix bugs in XFree86 because XFree86.org hasn't checked a patch into CVS. It could take 8-12 months for that to happen. Right now my XFree86 rpms have over 120 patches in them, and future builds will have more and more. If it didn't then what would be my job for example?
We don't really control upstream projects much, all we can do is make sure to send our fixes and whatnot upstream. But we certainly should never wait until upstream projects apply our fixes and release new versions of their software in order to ship it. That is just silly and needless lag. The distro would never be stable and/or would never ship.
Patches are and always will be needed to fix bugs.
And therein lies an essential role in providing a good distro - people who contribute to the stabilization of all the constituent parts, and the timely revision of relevant patch sets to produce viable packages as upstream changes occur.
A tip of the fedora to all of you maintainers who make quality releases possible. ;)
On Thu, 2003-10-09 at 17:27, David.Grudek@anixter.com wrote:
Currently they are all maintained by Redhat but since they plan on turning it over to the public more, that will be a harder task to watch what all these guys are doing to every package that comes out.
So, you think we're in total control of the disto? That we write every line of code in the hundreds of packages we ship? Let me tell you right now: We don't.
We ship free software. This software is written by thousands of contributors, many of them totally out of our control. Do you think that means that these packages will be unstable crap because we don't control them? No? Then why do you think someone helping us to package the software would?
Having the community working on a project does *not* mean that anyone can do whatever they want with it. This will be true for Fedora as much as for e.g. the Gnome project. When I get a patch for Nautilus (which I maintain in the Gnome project) from some random person I review it for correctness and send back comments, only if its good (and typically it takes a few tries to get it good) it gets committed to cvs. No unknown persons gets to fuck around with it however they want. If someone has shown a lot of sense and good technical quality I might let them commit their work directly to cvs, but these are generally people I know and trust well.
I find it amazing that people posting to this list, who obviously must be know about Linux has such a complete mistrust in the core ideas of Free Software. One of the basic ideas of free software is that if the community works together we can make better software. The reason it works is that, generally, competent experienced people maintain the software, and their main task is to say "no" to bad changes.
The belief that some company is better at coding/packaging than everyone else leads to something much like a proprietary system, but where the company also throws out the code when its done, for the users to look at. The power of Free software comes from the community participating in the development.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Alexander Larsson Red Hat, Inc alexl@redhat.com alla@lysator.liu.se He's a hate-fuelled small-town master criminal in a wheelchair. She's a strong-willed winged traffic cop on her way to prison for a murder she didn't commit. They fight crime!