Fedora's mid-life crisis [LWN.net]

The conversation started innocuously enough - or maybe it didn't. Rahul Sundaram's question was this: given recent decisions in the U.S. Supreme Court, might Fedora actually be able to point at repositories containing codecs which are said to infringe upon U.S. software patents? And, more to the point, regardless of what Red Hat's legal department says, does Fedora want to do such a thing? Fedora leader Max Spevack responded that, to answer this question, "the Fedora Board needs to reaffirm its larger strategy about Multimedia." There was some digression on how firmware does (or does not) differ from proprietary codecs. Then Mike McGrath broadened the scope further with a quick question:

What is our target market supposed to be?

The following is a quote from Bill Nottingham's response, but his message is worth reading in its entirety:

We don't have one! Seriously, I have yet to see anything that shows that we have a coherent market, a plan for attack, or *anything* along those lines.

So, we muddle along. Since no one has a plan or a target market, we implement whatever features the developers happen to think of, or random features vaguely relating to future enterprise development. Or we just incorporate the latest upstream....

Right now we don't have any overriding set of goals. So we never really say 'no, that isn't what we want Fedora to do' to anything that fits our simple 'uses open source, isn't completely targeted to obsolete things' mantra, and we attempt to do all of these things... which means we'll probably fail at all of them.

This message clearly resonated among the Fedora developers, none of whom stood up to say that he or she had a clear idea of who the target market is. Fedora hackers are looking over at Ubuntu, which has adopted a focused view of what it is trying to do and which has had significant success as a result. The Fedora project is seen as lacking that focus; it's not sure of what it's trying to do. As the distribution matures, its community is starting to ask itself some hard questions about where it is trying to go. It's a sort of free software project mid-life crisis.

Initially, Fedora's mission was seen - at least by outsiders - as serving as a proving ground for software destined to go into Red Hat Enterprise Linux and as a way to keep the venerable Red Hat Linux product around. So the target market will have been Red Hat itself, along with the Red Hat Linux users that Red Hat believed - almost certainly correctly - were an important part of making its enterprise offerings successful. There was no painful introspection in those days; Fedora mostly did what Red Hat wanted done - integrating Xen, for example - with the result that users began to despair of it ever being a truly community-oriented distribution.

The situation has since changed considerably. Red Hat still holds considerable sway over what Fedora does by virtue of paying a large number of engineers to work on it. But the distribution has become much more open and more driven by what its community wants it to be - should the community decide what that is.

There is a certain interest in turning Fedora into a polished desktop distribution. Doing so would require making some hard decisions: focusing on a single desktop, for example. It would require some sort of solution to the patent-encumbered codec problem. The support period - recently lengthened to just over one year - would probably have to be made longer yet. Much work would have to be done to make the various components of the distribution work together better; the tug-of-war between the two ways of configuring network interfaces (system-config-network and NetworkManager) was mentioned a few times.

Maybe, instead, Fedora wants to be a solid base upon which others can create finished distributions, much like the role Debian plays for Ubuntu. There is a certain amount of pride over the project's revisor tool which makes it easy to create derivative versions of Fedora. If this tool worked well with external repositories, others could take on the work (and legal risk, if any) of creating and distributing versions of Fedora with complete codec support, binary-only drivers, or any of the other things which are not consistent with Fedora's philosophy. Aside from the fact that Fedora is still seen (by its developers) as needing more "polish" to serve in this role, there is an interesting set of trademark issues which comes into play once a derivative distribution has something other than Fedora packages in it.

Fedora's trademark policy is already seen as an impediment by people making derived distributions (such as Dell's firmware updates live CD). It will be even harder for people trying to take Fedora into entirely new territory. The issues can be resolved by simply removing all references to the Fedora name, but there are advantages on both sides if derived distributions can claim to be based on Fedora. There has been some talk on how the policies could be changed, but anything concrete will happen some time from now, if ever.

Alternatively, Fedora could be a distribution for developers who want something close to the leading edge and who are less concerned with "polish." It's a legitimate audience, but it is also limited in size.

A number of other scenarios have been presented, but what is really required is for people to make the decisions and to get the work done to implement those decisions. It seems that Fedora is currently short of decision makers. Jesse Keating expressed it this way:

We seem to have a lot of sous chefs which are busy doing what they know, but no executive chefs with a grand vision of what will be on tomorrow's menus.

Anybody who aspires to be an executive chef can, if they actually try to make significant changes, expect a fair amount of resistance from elsewhere in the community. But perhaps the time has come for somebody who looks forward to that sort of challenge. The Fedora project has a solid base to build on and an increasingly open community process to help it get to where it wants to be. With the right focus on an interesting set of goals, Fedora could surprise the world. This distribution should have no trouble proving that it's not over the hill yet.

(Log in to post comments)

Fedora's mid-life crisis

Posted Jul 26, 2007 12:15 UTC (Thu) by rwmj (guest, #5474) [Link]

(Disclaimer: I work at Red Hat, but these are not the views
of Red Hat or Fedora, but solely my personal views).

You mention "RPM hell", but RPM hell (ie. trying to choose
compatible versions of RPMs that work together) hasn't existed
on Fedora for many years now. Fedora chose yum for package
management, in the same way that Debian chose apt to manage debs.

However, there are some very distinct problems with yum (speaking
myself as a former Debian & Ubuntu user).

(1) It's slow as hell. Operations which are near-instant
in apt such as solving dependencies take many minutes in yum.

(2) It gives out opaque error messages which don't help to solve
the actual problems.

On the other hand, RPM itself is a very sane package format compared
to debs, and creating RPMs is considerably easier IMHO (and I know
because I've created both on many occasions). Also creating yum
repos is much much simpler than the equivalent apt repositories
for Debian.

Fedora are working hard to fix the more egregious problems in yum.

It's important to note also that counting distros which use deb or
apt vs. rpm or yum isn't very helpful. Distros differ in many other
ways which means that just because they may happen to use (say)
deb files, does _not_ in any way mean that debs can be transferred
between them, any more than you can take an RPM from SLES and install
it on Fedora. In this sense, whether deb or RPM is "winning"
really doesn't matter at all. Whether particular distros are
winning mindshare on the other hand is more important.

Rich.

Package management

Posted Jul 26, 2007 14:19 UTC (Thu) by dowdle (subscriber, #659) [Link]

Well, I'd rather not see this topic get any farther into the package management wars... by contributing to this discussion... BUT... I do think your "many minutes" is an exaggeration. The only time I've had yum take more than a minute was when I was building an OS Template with OpenVZ... which is akin to a complete distro install. For normal, day to day operations, yum usually only takes a few seconds. I run my own repos and it is impressively fast sometimes... although yeah, apt would shave off even more of the install time. I think a lot of the lags that people have with yum are related to the mirror sites being slow... but I don't really have any data to support that.

Regarding the apt vs yum thing... and the term "rpm hell"... it is mostly about the packages themselves rather than the package manager. No one has really been relying directly on rpm for some time (just as Debian users don't use dpkg too often) with up2date, yum and others layered on top of rpm. I do give the Debian package maintainers credit as I think their packages have less problems... not that Fedora packages have lots of problems... but Debian has so many packages that Debian users don't go outside of the distribution provided packages as often as Fedora (or other rpm based distro) users do.

There was an article in LWN weekly a while back talking about some new package management system on some distribution, I forget which... but it was claiming that rpm development was dead because there hadn't been a lot of activity... but somehow he totally overlooked yum and the plugins for it that seem to be multiplying... but I didn't want to reply to that at the time because I didn't want to sound like I was an apologist. It just isn't cool when people use an article about one subject as a platform to advocate or slam another.

Ok, so if something I've said urks whoever reads this... don't take the bait... talk about the topics of the article rather than engaging in package management wars. :)

Package management

Posted Jul 28, 2007 22:10 UTC (Sat) by dberkholz (subscriber, #23346) [Link]

I wrote that article, and I was talking specifically about the RPM format. It's great to hear that higher-level managers like yum have an active ecosystem, but in the end, they're limited to what's expressible by RPMs.

Fedora's mid-life crisis

Posted Jul 26, 2007 16:19 UTC (Thu) by rsidd (subscriber, #2582) [Link]

Upgrades using yum are still not recommended on fedora, because of risk of breakage. On debian/ubuntu the worst that may happen is some third party packages may need manual removal and reinstallation. I've done apt-get dist-upgrade for 3 years, knoppix to debian sid to ubuntu, without breaking anything (though hotplug etc didn't work smoothly on ubuntu initially).

Taking more bait?

Posted Jul 26, 2007 17:57 UTC (Thu) by dowdle (subscriber, #659) [Link]

Have you seen the upgrade troubleshooting guide (or whatever it is called) for Debian 3.1 to 4.0? It goes on for pages and pages. I don't think that is a flaw... and is reasonable.

I have upgraded many minor versions of RHEL and CentOS with up2date and yum and have had no problems. Those distros have minor releases. Fedora does not. Fedora only has major releases. One way Fedora makes up for that is they have a ton, Ton, TON of updates during the lifecycle of a release... and yum is used for those.

Fedora has been much more aggressive with global system changes from version to version and with such a rapid release cycle... trying to make a flawless upgrade path from version to version would be a considerable headache... not impossible but would require a lot of work.

As previously mentioned, a typical Fedora user (probably) has more out-of-distro packages than a typical Debian user... putting more bumps in upgrade road.

The Fedora project has chosen to make yum based distro upgrades a low priority.

If doing a fresh install has a significantly less failure rate compared to a package manager based upgrade path (on those systems where it is available), some would argue that the fresh install option is the most prudent... but I realize it isn't for everyone.

For a home user, having /home on a separate partition and retaining that partition from fresh install to fresh install isn't seen as a problem... or at least I don't consider that a problem.

You have to also understand that even retaining stuff in /home can be problematic at times... especially when Gnome and KDE (and other apps) don't necessarily do a good job of migrating their older settings into later versions. Just because there weren't any errors during the upgrade and the system boots doesn't mean everything is fine. The only way a perfect upgrade system could be built would be if it were aware of all configuration files for all applications on the system on both a system and end user level, parse all of the config files, knowing all of the changes between versions, fixing all inconsitances... and then writing modified configurations. Debian's system tries to do that for the major server applications (which is good) but there is often an interactive dialog box that wants user input on which way to go (which is good)... but given the fact that there are way more apps where that process doesn't exist... you can see that there is a potential for problems. Debian does limit this to a certain degree by being less aggressive with version upgrades from release to release... and they have much longer release cycles.... so with regards to this, I think both distros have taken the upgrade path that makes the most sense for them.

Why did I take the bait?

Taking more bait?

Posted Jul 27, 2007 12:49 UTC (Fri) by rsidd (subscriber, #2582) [Link]

Why did I take the bait?

Actually, I didn't get any answer to my question: why would I consider switching from Debian/Ubuntu to Fedora, or recommend Fedora to a newcomer?

Perhaps rpm/yum is now as good, or nearly as good, as deb/apt. I doubt anyone would claim it's better. So what is better?

The question also applies to SuSE, Mandriva and others -- what do they offer? I can see what other distros do offer: Linspire and Xandros offer a Windows-like experience, Slackware offers a barebones experience, Gentoo is for ricers, the BSDs offer clean well-thought-out systems with readable source code and (depending on flavour) high performance / support for obscure hardware / paranoid security.

Why use distro X or Y?

Posted Jul 27, 2007 17:15 UTC (Fri) by dowdle (subscriber, #659) [Link]

I could simply reverse the question and ask, why would I want to switch from using Fedora to Debian or Ubuntu? It is all about choice and personal preference.

For me, I like Fedora because of the design, philosophy, project, and people...the rapid development/release cycle... and the fact that they update packages like mad during the lifecycle of a release. I like principles on using free software... and their unwillingness to compromise those.

I prefer using yum and rpm simply because I'm more familiar with them.

I like the various things I've stated before... about them willing to take more risks and use bleeding edge software (advancing development of them and maturing them faster by getting them into the hands of the user), advancing new technologies... like SELinux, Xen, KVM, etc. I like that Fedora is considered somewhat of a kernel testing platform because their kernels are usually very close to latest Linux kernel releases... and that gets the kernel tested by more users.

I like the fact that it is sponsored by Red Hat and all of the people that they employ to work on various GPLed products such as the Linux kernel, Xorg, gcc, etc. Red Hat really gives back to the community with the vast amount of development work they sponsor.

Those are just some reasons I came up with in short order so it isn't a comprehensive test. Perhaps they'd work for you, perhaps not.

For servers, I prefer RHEL and CentOS (more stable, longer support, more conservative, backed by the large server OEMS) and using Fedora on the desktop helps me get a glipse of what is coming... basically from every third release of Fedora.

There may be people who prefer the Fedora properties on a server, but I'm not one of them.

Bait taken

Posted Jul 28, 2007 8:46 UTC (Sat) by man_ls (subscriber, #15091) [Link]

Your message reads like a lot of excuses about why Fedora upgrades do not work and why it is so hard to do. But it sidesteps the real issue, which is that in Debian and in Ubuntu it works, so it can be done both by volunteers and by paid professionals. And Ubuntu doesn't have long release cycles. And Debian has major updates.

Sure, it is not a trivial problem, and every time I upgrade my debian-based machines I am amazed that it works. And it is probably a lot of work. We already know that. But it works, while in Fedora people are complaining about it. To me it looks like poor engineering, and it reflects badly on Red Hat.

Bait taken

Posted Aug 2, 2007 7:38 UTC (Thu) by yeti-dn (guest, #46560) [Link]

> in Debian and in Ubuntu it works

I'm sorry, this is just propaganda. I have upgraded between Fedora releases with yum update many times, and it always worked, possibly with a few packages requiring manual handling. Whereas when I upgraded Ubuntu, it was one-click, proceeded smoothly -- and the system did not boot afterwards. So, as everything, it works for someone, sometimes.

Bait taken

Posted Aug 2, 2007 8:38 UTC (Thu) by man_ls (subscriber, #15091) [Link]

It is not just propaganda; at least dowdle said above:

The Fedora project has chosen to make yum based distro upgrades a low priority.

And that does not happen in Debian land. Of course it will have flaws, but if it does not work, it is considered a serious bug.

Maybe that is not the official position anyway, and that would be nice. I don't know, I have never installed or upgraded Fedora or RH.

Bait taken

Posted Aug 2, 2007 14:33 UTC (Thu) by dowdle (subscriber, #659) [Link]

man_ls,

I am not am not an official Fedora representative... just a user and community member since like... Red Hat 3.

Just ran across a posting by Kevin Fenzi on the Fedora People site where he talks about his recent experiences upgrading three machines from FC6 to F7 via yum... so I thought I'd pop back here and post it and I see that there have been a couple of additional comments added since I last checked.

Yeah, I considered mentioned the fact that (according to one source I read a while ago) the failure rate for one Ubuntu upgrade was 40%... but I didn't previously mention it for two reasons:

1) I didn't feel like hunting down the article I read

2) It was from release before last I think

3) Unlike you, I don't get pleasure from ripping on things

4) Since you've been in the Ubuntu/Debian community, I'm sure you were familiar with it already... and I didn't want to have to listen to your reason it failed... because it is a past release anyway... and who cares... they've done better since and will continue to do so

You seem to think that if Fedora doesn't recommend yum based upgrades that it is a major flaw... and say that the Fedora community is clammering for it. Hmmm, I'm like... a member of the Fedora community and I really haven't heard a lot of clammering. I've considered doing yum based upgrades (because it is possible) but the last few releases, I've not bothered... because doing a fresh install is a lot quicker and easier for me (can't speak for others).

Oh, almost forgot... here's Kevin's posting about his recent upgrading via yum:

http://www.tummy.com/journals/entries/kevin_20070801_113809

Whoopie. Who cares... some people do like to upgrade via yum.

I do think it is rude of you to invade this topic with your negative message. I do NOT go into Ubuntu related stories and talk bad about it... but there seems to be a pattern where Ubuntu people invade Fedora topics and post anti-Fedora stuff. Oh well, I'll get over it.

Bait taken

Posted Aug 2, 2007 15:24 UTC (Thu) by man_ls (subscriber, #15091) [Link]

Sorry if I have bothered you, I can assure you that it was not my intention in the least. I don't expect anything at all from any distribution that I may or may not use; I don't even pay for them, so I am grateful for what I receive in any case.

Bait taken

Posted Aug 3, 2007 1:32 UTC (Fri) by dkite (guest, #4577) [Link]

Thank you for exposing the Ubuntu propaganda. I have a metric
called 'mean time to command prompt' when installing or upgrading a
distro. Ubuntu was second to Gentoo.

Suse was quite good, although the end result was quite heavy. And the rpm
management was glacial. I didn't try Redhat or Fedora.

I'm not suggesting that others have not had easy or flawless upgrades. I
consider myself outside of the Ubuntu user profile. Obviously, because
they wouldn't have released it if my experience was common.

Derek

Fedora is quite bloated

Posted Jul 26, 2007 14:49 UTC (Thu) by walters (subscriber, #7396) [Link]

Is being a base for embedded systems or appliances a goal for Fedora? Not explicitly. Could it be? Sure. If you want to see it, you could step up and make some proposals for how it could happen.

Some of the things you mention like glibc-common are basically just rm -rf in a post-image script.

It seems sane to me to run at least D-BUS on an embedded/appliance system unless you're *really* small.

SELinux would be a recompile of a few packages probably.

Anyways, the point is unless someone steps up and takes leadership on how Fedora could evolve to serve embedded systems (and there has been a few people interested), it's not going to just magically happen.

Why Fedora then?

Posted Jul 27, 2007 12:13 UTC (Fri) by walles (subscriber, #954) [Link]

If Fedora has all those problems, why do you still use it?

try ALT Linux then?

Posted Aug 2, 2007 21:17 UTC (Thu) by gvy (guest, #11981) [Link]

Well, you can look at PLD or ALT as a technocratic/meritocratic examples of RPM-based Linux distros. ALT is historically known for being very solid base, ApplianceWare NAS was built on it since ~2001.

I guess early introduction of more high-level PM tools than bare RPM helped both of them, along with more or less well-shaped goals (in ALT's case, reasonable security/usability mix among them).

There's Server release and a Desktop RC2 (20070801) for everyone's reviewing and using pleasure; there's also 4.0/branch on FTP including spt tool which is used to create those ISO images and OpenVZ template cache as well (BTW ALT's are by far the smallest of everything in that area of download.openvz.org).

The catch is the same as with nginx: it's a Russian-spoken project, mainly. But for those who decide to delve in such projects (like Ruby, a Japanese-spoken project ten years ago) tend to discover beautiful things.

And yes, there are no concerns on software patents in Russia so every driver and codec we could enable or include within copyright is there :)

PS: I hope that seeing Fedora on our tracks is good for them, even if it's 2--3 years behind in different aspects (like this one). It's always better to raise problems and solve them than to ignore them!

Fedora's mid-life crisis

Posted Jul 26, 2007 14:57 UTC (Thu) by dowdle (subscriber, #659) [Link]

While I do appreciate hearing the Fedora developers talking and planning on how to improve Fedora... I think some of it is motivated, as were some of the changes in Fedora 7, by envy of Ubuntu's supposed popularity. Now before I get flamed by an Ubuntu fanboi, let me say that Ubuntu is a fine distribution and although I prefer Fedora for personal desktop use, I have actually recommend Ubuntu to some users that I thought it was more appropriate for. Of course I also recommend Fedora too... but yes, Fedora is better for users who have some Linux experience.

I don't mind that Fedora doesn't have a "target" but do think it would benefit if the developers do start up a few more projects (as long as something else doesn't suffer as a result) and put additional focus on improving things in various areas. What areas? I don't really care... as I'm fairly happy with the status quo.

The whole notion that Fedora should drop a desktop or two and focus on one... is silly. Look at Ubuntu... and Kubuntu... and Xubuntu... and all of the rest. People want choice. I run lab machines in an academic environment and if I had to limit my users' choice to one particular destkop environment there would be no end to the bitching and I'd have to change distros... or go outside of the stock distro packages. I don't use Fedora in my labs because I prefer a longer support cycle... but I do use something that inherits from it.

Which users want a desktop where all of the decisions have already been made for them? I'm sure there are some, especially newbies... but looking at my users... I'd guess more than half of them spend a considerable amount of time customizing their desktop environment because they don't want to look like everyone elses.

Asking Fedora to narrow its scope is akin to asking the Linux kernel developers to pick a particular target. It just doesn't work that way. I forget who said it first, but Linux is like a liquid that flows into every crack and space it can find... and I like Fedora because is isn't focused on any particular thing... and does several things well... Jack of many trades... or pick-your-adventure as some of the Fedora developers were saying. Yes, it is a lot of work being so broad.

In a lot of ways, Fedora and Debian are a lot alike... and I'd prefer to see Fedora flowing in more directions until it has as large or larger of a package base as Debian. That might not be as marketing friendly as distro talking points such as "polish" and "integration" or having clearly defined "targets"... but it is a big reason I continue to like Fedora.

I'm so glad so many good distributions exist... because it creates a lot of competition between them... which is so good for the end users.

The Fedora developers should also feel good about the pushing they do to advance new technologies... and new applications... ahead of most everyone else. That is another area where they stand out. Red Hat and Fedora give so much back to the Linux distro pool with their continued support of developers in many major projects and I just don't see that in many other distributions.

The Fedora developers also seem less afraid to take risks by adopting things early and allowing them to be used by the userbase before they are prefect... so the userbase gives more back to the development cycle with real world usage, testing, and bug reporting... leading to faster product maturity. In that they have inherited from the somewhat notorious Red Hat Linux .0 distro releases before the birth of RHEL.