Submitted by Max Spevack, Fedora Project leader posted a outline and expected impact of the Fedora 7 release: “One of the Fedora Project’s success metrics is building and running itself in a way such that no single entity can completely control Fedora’s fate. Fedora 7 gets us there, insofar as there is no “secret sauce” in the ability to spin a Fedora distribution. Nothing is hidden.”
Fedora 7: What, When, and Why
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Subtle jab at Gentoo 2007.0 “Secret Sauce”?
Ha ha. No. Nothing to do with codenames for another distribution.
Fedora 7 drops some features but not all of them.
Even if Fedora sucked, I’d still use it because of it’s security features that most other distributions lack.
But quite frankly, Fedora has been completely stable, quite easy to use and able to do everything all other distributions can. I really see no need for me to use anything else.
“””
Subtle jab at Gentoo 2007.0 “Secret Sauce”?
“””
Didn’t secret sauces get pretty much killed by the litigious climate today?
Imagine the family taking Grandma out for dinner at Colonel Sanders’. The 11 secret herbs and spices happen to include coriander, to which Grandma is allergic. Grandma takes a bite, goes into anaphylactic shock, and dies horribly in front of her daughter, grandchildren, and the new great grandchild. (The contorted look on her face as she lies there dead next to the lavatory door is too disturbing to describe here.)
I wonder who gets sued for millions of dollars?
Even more than that after the guy walking out of the lavatory accidentally hits Grandma in the head with the door on his way out.
OK. Totally off topic. But this was the little vignette that played out in my (admittedly twisted) mind when I read your post.
Anyway, it seems like secret sauces, software or otherwise, are an idea whose time has come and gone.
It’s not meant to be a jab at Gentoo in the slightest. Purely a coincidence.
I was looking the feature set, and I don’t want to sound anti-Fedora, but it is pretty bad by the fact that the new startup/shutdown has not be merged into Fedora (been classified as ‘OUT’ in the feature matrix for Fedora 7).
Personally, having had a look through all those features which have been excluded, one has to ask – why even move to Fedora 7 or even use it? So far, according to the wiki, these are not going to be in there (http://fedoraproject.org/wiki/Releases/7/FeatureList):
Releases/FeatureBootShutdownSpeedup (OUT)
Releases/FeatureCodecBuddy (UNCERTAIN)
Releases/FeatureFedoraDirectoryServer (OUT)
Releases/FeatureFixWakeups (OUT)
Releases/FeatureNewBuildSystem (OUT)
Releases/FeatureRPMYumEnhancements (UNCERTAIN)
Releases/FeatureRandR1.2 (UNCERTAIN)
Releases/FeatureRockSolidWireless (UNCERTAIN)
Releases/FeatureSyslogNG (OUT)
Releases/FeatureTexLive (OUT)
I’m sorry to say this, but its terrible; quite frankly, they should push back release another 2 months and make Fedora actually worth upgrading to or otherwise there is little reason to use it when compared to what else is out there.
Thanks for the list. There are many cool IN things there .
They should just change the release schedule to 9 months. Because this 6 months or whatever it is, is not working.
“I was looking the feature set, and I don’t want to sound anti-Fedora, but it is pretty bad by the fact that the new startup/shutdown has not be merged into Fedora (been classified as ‘OUT’ in the feature matrix for Fedora 7). ”
It was clearly marked as a feature that needed a owner even when it was announced so I don’t think anyone who paid attention would have expected it to go in immediately. Switching a init system is not something a distribution can take lightly.
“Personally, having had a look through all those features which have been excluded, one has to ask – why even move to Fedora 7”
It has about 20 out of 30 features originally planned which was very ambitious so this is actually very good.
This release is really about infrastructure improvements aimed at providing better access to community volunteer contributors more any particular feature as the announcement has highlighted. There are indeed significant new functionality available now.
A short overview is at http://lwn.net/Articles/231862/. Delaying a release for something like 2 months as you suggest doesn’t work as it frustrates developers with a longer freeze, other including software gets delayed further and there will be always more improvements expected.
If you push yourself to do a more regular release interval and if you don’t get certain features in the next release is always just around the corner 6 months ahead. Collectively though the improvements will add up quickly over a short period of time.
Why create a new one at all? why not use Apple’s one they created? or a re-implement the Sun one? or how about just using the one Ubuntu uses? why do people in the Linux distribution world feel the need to re-invent the wheel when there are already solutions in place?
Regarding the 6 month cycle; personally, it is too quick – their aim should have been to get all the core done by the time 2.18.3 of GNOME was available, then merge it in, after most of the big bugs had been hammered out of GNOME. Sure, it would have resulted in a ‘out of date’ GNOME, but the over all distribution would be alot more stable.
Edited 2007-05-09 03:05
“Why create a new one at all? ”
Who said anything about *creating* a new init system? I said switching to a new init system. If had bothered to read the specification before commenting you would have found out that all the other init systems you mention are described as alternatives that need to be evaluated and precisely because there are many alternatives you need to be careful about what option you pick.
“Regarding the 6 month cycle; personally, it is too quick”
Well you are not involved in the distribution so your personal judgement isn’t going to make much of a difference. You can’t just align yourself with GNOME. There are a number of other components that are important that varies from release to release. For this release the major difference is in the infrastructure.
Switching to a completely new open and external build system and merge Fedora Core and Extras is a HUGE amount of effort. You won’t see the effects immediately but it increases the potential for contributors considerably.
6 months is kinda the standard for many distributions these days. Originally starting with Red Hat Linux has grown into Mandriva, Ubuntu etc.
With 6 months intervals you don’t wait for many features but keep pushing out incremental improvements.
Edited 2007-05-09 03:12
I can and will forgive Fedora for not being as “inovataive” as it may first appear for one simple fact.
The amount of work that it has taken to do the merge between Fedora Core and Fedora Extra’s, was monumental.
Every single package from both the redhat folk and the extra’s folk had to be RE-checked, each spec file had to be reviewed. I am sad to say, I intended to help out with this, but after seeing the work required, and the depth of knowledge required (and as such, reducing the number of people that COULD assist) it was just a bit out of my league.
I am hoping (as are many) that once this merge has completed things should start to move faster (though, I’m fairly happy with the current rate of progression to be honest with you).
if one had of read, these Features are not even stable enough to even put them in a distribution , , the Bootup/Shutdown improvement comes from Ubuntu which is unstable even in Ubuntu so why would fedora/Redhat use such an unstable thing like that?
they all should be in Fedora 8, depending if there stable enough
Because Fedora is free as in freedom? And the developer community cares about this kind of freedom? Just a thought. BTW, your paste of the list is outdated.
Fedora will continue to be a time-based release. We try to do a release every 6 months, and what that really means is that by the time we deal with inevitable slips and delays, hopefully we’ll get a release out more or less every 7 months.
In terms of features — I think you will continue to see us have lists of “features that we’d like to get into Fedora” but for the most part the features that make it in or get left out will be decided based on when they are ready.
Fedora 7 was an exception — the Core/Extras merge, along with the opening of the build system (koji) and compose tools (pungi) were considered Must Have features no matter what — we would hold the release indefinitely for them.
But something like improving the boot/shutdown time — those are things that we want to get into Fedora, but if they aren’t ready, we won’t hold the release for them.
The way we really need to lay out or wiki is to simply have a section of “Self Contained Features That Can Be Added to a Fedora Release” and be able to track their development status independent of any particular Fedora release. Once a feature is ready, it goes into Rawhide.
I have thought for a long time that Fedora was going in the wrong direction, and this just adds to my belief that Fedora has lost its way. My choice has always been Suse, either Novell or OpenSuse. I wish Fedora would take a page from their book. Even better, I wish Fedora would push to have their software repositories as well kept as Suse’s Yast/Yum are.
The release a new version just because it will be new is simply not enough. There should be a valid and good enough reason, and I simply do not see this with Fedora
“Even better, I wish Fedora would push to have their software repositories as well kept as Suse’s Yast/Yum are. ”
The recent merge of what were once separate Core and Extras repositories into what is now the Fedora Package Collection means that the entire Fedora community – not just several select Red Hat employees – has much deeper and more direct control over the flow and maintainance of Fedora’s repository and packages. With this merge comes official QA reviews of EVERY package that was once in Core. This should significantly increase both the quality and quantity of packages available by default.
—
“The release a new version just because it will be new is simply not enough. There should be a valid and good enough reason, and I simply do not see this with Fedora”
No, the main user-visible changes in this release are newer kernel/drivers, the Flying-High artwork, and updates to the GNOME 2.18 stack. The main goals of this release, which are mostly being met, are the repository merge as well as the infrastructure for the new build system (Koji) and LiveCD spins (Kadischi). As it stands, these goals are much more developer-driven than new themes or new drivers. Hopefully, this new infrastructure will make Fedora 8 and further much better, though!
you use Suse? may not for much longer if more employee’s leave Novel ,
atleast Redhat hasn’t signed those dummy papers that Novel did, Redhat will grow stronger from Novels mistakes , Fedora Might have its issues with repositeries but sure is more stable IMO than Ubuntu 7.04 or OpenSuse looks at How Many problms 10.1 had ( OpenSuse ) OpenSuse May aswell be renamed to MicroSuse
how cares if novell employees leave. the community distro will still be around.
That is an aweful comparison to make, with OpenSuSe and Novell.
Those distro’s load up with MONO crap and are encumbered with agreements legal and financial with Microsoft.
Pick ANY distro to compare Fedora’s problems with BUT THAT ONE.
Seeesh.
-Hack
I think it is funny that every time there is a Fedora topic posted the Fedora bashers come out in groves. Some complain about what didn’t make it in. Some complain about schedules not working. Fedora is more about putting together the latest Gnome, Kernel, and applications and packaging it up into a usable distro. So everything didn’t make it in. Big Deal. I am using Test 4 to write this. I find it faster and more stable than version 6. Seems like improvement to me. I just find it odd that people stick to regurgitating old perceived deficiencies when the article linked to is about the open process used to build Fedora releases. I am not trying to be insulting but it would be nice to actually see a discussion about the actual topic being discussed in the article without the tired old, “but it is doing this wrong, and this, and I want it to be like Debian”, yawn….
I used to use fedora before I switched to Ubuntu. I’ve been thinking about installing it on one of my spare hard drives. I’ve always liked fedora but like apt-get/synaptic more.
This is from the RPMForge repo (on CentOS). It’ll surely be there for Fedora 7 too.
I’ve never felt the urge to use it, but if you like Fedora and prefer apt-get/Synaptic, voilà.
Hmm, in my experience using apt4rpm on Suse, it ended up being a matter of either using the built-in tools or the apt, but not both; otherwise YaST would get very confused and I had to lock all of the apt-installed packages away from YaST’s meddling. Not sure if that’s as much the case when using apt in conjunction with yum on Fedora.
Aside from that though, the biggest advantage to using a Debian-based distro compared to RPM is that dpkg is simply faster. At least in my experience. Again, maybe this has changed in recent releases of Fedora. If so please let me know.
Yum performance in Fedora 7t4 has been improved.
I don’t use .deb-distros at the moment, so I can’t compare. I can compare it to FC5/6, and CentOS 5.
However, the Yum command line utility has always performed well for me anyway. Those few minutes (?) of supposed extra performance of dpkg/apt.. I always wonder what people are doing while installing packages. Staring at the screen till it’s done?
And how many times a day do we actually install something?
I always wonder what people are doing while installing packages. Staring at the screen till it’s done?
And how many times a day do we actually install something?
Well, why not ? Pesonally I will eagerly await Full spins both i386 + x86_64, shipping from local seller. I expct prices near 1/3 of my month revenew (dont even ask how much it will cost here if I try download from ‘net). So, imagine day i return to home with big smile and small package with DVD boxes with broken a bit flexyglass corners diring shipping. First I want to sure that at least one spin (i386 or x86_64) inslallable. Not bit-to-bit exact with original using bootup md5 test, but just have unbroken subset of packages that i can see shell and do the rest install by hands replacing broken rpm’s from previous version or something like wodoo-magic. So I insert DVD and keep staring to full install messages during setup. Stupid? maybe. For guys that live in Moscow with $18 per-month unlimit connection, and enormous high revenue (~2x for equal work and more) yes. But for me, if DVD have scratch, it mean another ~3 weeks delay+money lost. It is not fun, and if you know that you obviosly pay attention to yum install performance, delays, swapping storm during install and unexpected >4 hr standard upgrade in 128 mb box ( 4 hrs was enough for me I was reset box during FC4-FC5 upgrade, cry on my old setup and do full repartition+install).
Conclusion: yum/rpm performance is matter. It make first impression of distro quality. It stress disk subsystem and work like first ugly OS benchmark. If RedHat want to increase Fedora user base ( and RHEL as well) then yum/rpm performance and RAM usage oprimisation are important.
Yes, well… sometimes I do stare at the screen
…. I guess mostly it matters 1) on initial install and package selection and 2) when it comes to installing updates and the like…. Anyway, I’m thankful for any minutes that I gain being able to use my computer instead of just waiting around for it. But I’m happy to hear that they’re optimizing yum. I’m actually pretty excited about F7, despite my having written off Fedora in favor of other distros a couple of times already.
Edited 2007-05-09 22:07 UTC
I really do like the approach that fedora has taken regarding open source only. At least they are not feeding
MS fud. I predict that the merge of core and extras will not only increase the quantity of packages but also increase stability end reduce problems with dependencies.
Wow, to me this pitch about the whole open distro-creation, anyone-can-do-it process sounds like Fedora is just waiting and wishing for some third party to come along and be to Fedora what Ubuntu is to Debian: a subset of the base distro, built for the masses. Sounds like a good idea to me. Maybe someone will jump on it!
Please, not yet another distro.
A Fedora spin-off built for the masses?
You’ve got a couple hundred million dollars in your pockets like Shuttleworth?
Anyway, that won’t be necessary. Fedora 7 will be launched (a.o.) as an installable Gnome live-cd, and as an installable KDE live-cd. I tested them and they work quite well. That looks very beginner-friendly to me.
Please, not yet another distro.
A Fedora spin-off built for the masses?
What, you don’t want to see Kedora, Xedora and Edudora?
http://download.fedora.redhat.com/pub/fedora/linux/extras/6/i386/re…
http://download.fedora.redhat.com/pub/fedora/linux/extras/6/i386/re…
The list of changes actually looks pretty good, I have always admired fedora and used it for years before switching to gentoo, it just worked, and its stability (on my machine) was far superior to ubuntu and suse (suse’s yast was always a nightmare for me, and never actually changed what I wanted to change, leaving the system unbootable many times. although this may have improved in recent releases). I look forward to using fedora on my home server in the near future, and recomend it to anyone looking for an easy distro with an excellent community there for you when you want to start toying around (eg adding multimedia support etc..).
To really gauge the strength of Fedora, you have to look beyond what packages is does or doesn’t have. I am not a Fedora user (I use Ubuntu), but the LUG I attend is mostly Fedora users. The strength of Fedora truly is the community. By community, I mean more than just forums, etc. Fedora has a vibrant developer and user community. It has tremendous grassroots strength. While I may not prefer the actual delivered product, I cannot do anything but admire the community around Fedora. The srength of Fedora is both a strong central presence, as well as a strong community presence world-wide.
This Ubuntu user says, “Go Fedora!”.
by one week – such an announcement has just been posted to the mailing list. It’s due to the extras/core merge that needs little more time to be ironed out.
In my opinion, the best things in Fedora, when compared to other distros, are related to security (and server use):
http://fedoraproject.org/wiki/Security/Features
None of the other major distros (those listed, for example here: http://distrowatch.com/dwres.php?resource=major ) in their default install seem to achieve the same level of security as Fedora (although – I’m not sure if SELinux – probably the most talked about security feature in Fedora/Redhat nowadays – is the ultimate best way to go in improving Linux security? See, for example here: http://www.grsecurity.net/lsm.php ).
If we forget the top level server security, and other server features for a moment, and consider desktop users’ needs instead, I think that, for example, Ubuntu, or Novel’s OpenSuse, are probably better, and more advanced desktop operating systems than Fedora (although the differencies between those three may not be very big in the end).
Edited 2007-05-09 16:28
I don’t get the argument that Fedora doesn’t innovate.
Pretty much any of the major desktop technologies have developers who work on Fedora. DBus, HAL, GTK, PolicyKit not to mention the scary productive kernel hacker Ingo Molnar also does development on Fedora. Red Hat funds maintance of large parts of what we call the Linux system and that is done via Fedora.
The new GNOME Online Desktop project.. yes done within the realms of Fedora. Fedora is also the first distro to ship with ConsoleKit and PolicyKit and much more. SELinux is largely being developed with Fedora as the reference platform and we are still the only major distro to ship SELinux enabled by default along with a lot of other security measures.
Fedora is the base of the most interesting Linux deployments from a technology POV: Yellowdogs linux for the PS3, OLPC. Red Hat builds their enterprise product on Fedora technology, by extension Red Flag, the worlds most used distro, builds on Red Hat’s products so that’s also containing Fedora packages. Naturally Fedora itself is also a superb platform but given it’s nature it might not be suitable for certain deployments.
As you can tell I have Fedora on my machine, it’s not so much fanboying (we aren’t without flaws, nobody is) but I certainly don’t get the lack of innovation argument. There’s a lot we can deserve a swift kick for but innovation certainly can’t be one.
“””
I don’t get the argument that Fedora doesn’t innovate.
“””
I see that you have fallen into the trap of thinking that the word “innovate” has any real meaning today.
MS pretty much ruined any utility that the word may ever have had.
Best to just get on with *doing cool stuff* and forget about debating “innovation”.
Edited 2007-05-09 22:11
For speeding up Yum on older versions of RH systems and how that works, see (a.o.):
http://blog.danieldk.org/post/2007/05/10/yum-metadata-parser-for-yu…