Linux Desktop Distro Shootout

An anonymous reader writes "InfoWeek has posted an open-source OS comparison. Linux Shootout: 7 Desktop Distros Compared pits openSUSE, Ubuntu 8.4, PCLinuxOS, Mandriva Linux One, Fedora, SimplyMEPIS, and CentOS 5.1 against each other. And the winner is ... Ubuntu. Author Serdar Yegulalp writes: 'Ubuntu 8.4 remains one of the best desktop distributions for many good reasons: it works with almost any hardware you throw at it, and has tons of features for both existing Linux users and prospective converts from Windows.' He also gave openSUSE points for ease of use on the desktop, and Mandriva kudos for ease of administration."

Comments

8.4?

[Anonymous Coward]

8.04.

Re:8.4?

[doti]

Isn't it unfair to compare Ubuntu 8.04 to Fedora 8?

Fedora 9 will be launched soon, they could have used the beta.

Fedora 8 could be compared to Ubuntu 7.10

Re:8.4?

[timeOday]

If they had tested a beta and encountered hiccups, people would have complained that evaluating a beta against a production release isn't fair. And between 7 distros, I'd guess at least one has an upcoming release at any given time.

Re:8.4?

[aaronmarks]

With that logic though, they shouldn't have tested Ubuntu 8.04 because they apparently tested that as well while it was still in beta.

why CentOS?

[trybywrench]

Isn't CentOS the free version of Redhat Enterprise Linux? Why is it in a desktop linux shootout?

Re:why CentOS?

[lgarner]

It includes elements from all RHEL packages: RHES, RHAS and RHED.

Add free version.

[Anonymous Coward]

Here

Re:Add free version.

[pthisis]

"couldn't care less" not "could,"

And "begs the question" doesn't mean "raises the question" or "ducks the issue". It refers to a specific form of argument which _does_, in fact, attempt to answer the question--but does so by assuming the conclusion in one of the premises. Specifically it's an argument of the form

p implies q
suppose p
-> q

Where "suppose p" really is "suppose my side of what we're arguing about is true".

Better methods of statistical comparison.

[Thanshin]

1 - Make a live disk of each.
2 - Build many identical robots.
3 - The round starts upon insertion of the disk.

Last standing robot wins.

THUNDERDOME!!!

Hey!!!

[morgan_greywolf]

No fair! Who gave OpenSUSE that AK47?!

Ubuntu 8.04

[SkankinMonkey]

I'm quickly finding that I prefer 7.10 to 8.04. The overall system seems a lot more bogged down, lots of freezes with programs that never occurred in earlier versions. I do like a lot of the new functionality but I hope that they iron out some of the outstanding issues (especially considering it's supposed to be a LTR).

Re:Ubuntu 8.04

[Anonymous Coward]

Feel free to file a bug: https://launchpad.net/ubuntu/+filebug

Re:Ubuntu 8.04

[SkankinMonkey]

I've submitted almost all of my crash reports to them, and all the crashes I've experienced are known bugs (and had multiple page threads on their forums during alpha/beta testing). That being said, I think they should have held off a bit on a final release and squashed a few more bugs that were pretty proliferate and user inhibiting.

Re:Ubuntu 8.04

[TheLink]

I heard the freezes are due to some scheduler thing they did - rather than all processes competing equally, you have some weird situation where programs that have root and user instances have problems with one starving out the other.

Re:Ubuntu 8.04

[Dystopian Rebel]

There was a decision to use the old scheduler on the Desktop version of Heron. It is causing problems. Try the Server version.

I find that it's as wise to wait for stability in an Ubuntu release as it is with an MS Windows release. The difference is that stability comes to Ubuntu faster. (o:

I will give Heron a month or two to settle down and then switch.

Ubuntu does more right than any other Linux distribution ever has.

Re:Ubuntu 8.04

[twistedsymphony]

Release Schedules are nice and all but what's the point of bundling up a "stable release" if it's not actually stable?

If you want to download the latest SVN snapshot every 6 months that should be your prerogative but I've been burned too many times by "stable release"s that weren't actually as advertised simply because someone said "it's release day... SHIP IT!".

I always do some form of testing but it's a lot of wasted effort if you're installing something that you assume is already as clean as it can be, and it's really not.

Re:Ubuntu 8.04

[nine-times]

I guess I just figure that a lot of "stable" software won't really have all the kinks ironed out until after release. When something is released, it's probably going to put onto hardware that no one was testing on, and it's probably going to be used in ways that it wasn't used during testing.

I agree that if there are known major bugs that will be extremely common, or bugs that are show-stoppers (e.g. cause significant data loss), then release should be pushed back. But if you want something extremely stable, then you might consider holding back your upgrade for a little while.

But I'm not making an argument from principle. I'm just saying that, from experience, I've never seen anyone get something 100% bug-free. Even Debian stable can have some quirks. So I'd rather have a regular release schedule than have progress on Ubuntu held back until every little bug can be worked out.

And the winner is .....

[Taco Cowboy]

No matter which distro takes the #1 spot, the real grand prize winner is ....

                                                              THE USER !!!!

Fedora

[BountyX]

Fedora 9 comes out 8 days 3

Re:Fedora

[pavon]

Fedora 9 comes out 8 days <3 Fixed that for you.

"Almost any hardware you throw at it"

[loteck]

I sincerely enjoy the Linux experience and appreciate the community, but this statement is positively absurd. Ubuntu's own help files contain extensive lists of wireless cards that have a big fat "No" listed under the "Works out of the box" column. And that's just wireless cards.

One of the primary reasons that the average person abandons Linux is the frustration caused by these types of misleading claims. Somebody says, "Hey, virtually everything works out of the box!" and they think... wow, well, I buy my stuff at top retailers from top brands, surely then my stuff is supported.

Unfortunately for them, their stuff may not work at all, or may work partially. Lots of gotchas for Video cards, scanners.. the list goes on and on. Nobody is well served by making statements that indicate anything except that hardware support is still a major obstacle for the adoption of Linux on the desktop.

Re:"Almost any hardware you throw at it"

[Waffle Iron]

In keeping with industry practices, maybe they should have rephrased the claim to say: Almost any random hardware is "Ubuntu Capable".

Re:"Almost any hardware you throw at it"

[strabes]

I feel like everyone on slashdot should know this, but I'll repeat it once again. Not having support for various wireless/video/etc cards is not the fault of linux or the kernel developers. It is the fault of the vendors for not providing proper drivers and/or documentation. This will only improve with time as the popularity of linux grows and greater pressure is put on vendors to provide the aforementioned drivers & documentation.

Re:"Almost any hardware you throw at it"

[Hatta]

What kind of weird ass hardware are you using? The reason most people say that most hardware is well supported on linux is because it's the truth. If you're that unlucky that you bought oddball hardware, that's too bad and it does need to be fixed. But it really does work with almost all hardware you throw at it. Emphasis on the almost.

And I'll point out that OS X works with even less hardware than Ubuntu does. That didn't stop you from choosing it. Why should hardware support stop anyone from choosing Ubuntu?

Re:"Almost any hardware you throw at it"

[Kaitnieks]

I recently installed Ubuntu on laptop (my 1st linux) and had problems with wireless as well. Thing is, the wireless card problem can be easily fixed, but for some strange all tutorials, manuals and forum posts offer long, non-working instruction lists, that involve "wget" and "make". It's like asking - please, go away. The real solution was to open synaptic package manager, configure it to use 3rd party repository and install ndiswrapper. That's it! Ndiswrapper found and downloaded the drivers automatically and everything was bright and sunny again. Why isn't ndiswrapper in standard installation I have no idea.

Re:"Almost any hardware you throw at it"

[shutdown -p now]

I feel like everyone on slashdot should know this, but I'll repeat it once again. Not having support for various wireless/video/etc cards is not the fault of linux or the kernel developers. It is the fault of the vendors for not providing proper drivers and/or documentation.

I feel like everyone on Slashdot should know this, but I'll repeat it once again. Users, for the most part, don't care why something is not supported - if it isn't, they are simply not going to bother with that particular distro/OS. Blaming vendors (even when fair) does not achieve anything - they just shrug and say, "what do we care about your niche geek OS?", and users get even stronger impression that they should stay away from that weird Linux thingy.

And, no, I don't know a solution to this short of waiting and hoping for the better. But we certainly shouldn't be telling people that "most hardware works in Linux" - because that is outright lie.

Re:"Almost any hardware you throw at it"

[Necreia]

The "Average User" doesn't install their OS. It comes preloaded and configured from Walmart / Dell / etc.

How many of those distros

[pembo13]

even claim to be a desktop distro? I use Fedora on my desktop, but I don't think they claim it to be a desktop distro.

Why Mandriva One?

[CastrTroy]

Why did they opt to use Mandriva One, over Mandriva Free? Mandriva Free is a bigger download, but comes with a lot more software on the disk. It also seems more suited to an actual install, whereas Mandriva One is more of a Live CD.

Lastest Ubuntu, Older other distros

[vossman77]

Fedora 9 comes out in two weeks, but we are comparing the nw born Ubuntu to 6 month old distros. Ugh. Let's compare apples to apples people!

Recent Linux convert from Windows

[goltzc]

I need to start out by saying that I am a web developer and other than very basic work deploying code to HPUX boxes at work I have had 0 exposure and no formal training with *nix Operating systems. I started playing around with Ubuntu during the Feisty release on my windows desktop with Wubi. Once I saw that I could get all my work done reliably and how stable Ubuntu was I knew that It would be my main operating system... someday. When I purchased a laptop with Vista preloaded on it I realized that Ubuntu was going to have to come to the rescue sooner than I was planning. Right about the time gutsy came out I put in on my no frills middle of the road laptop and haven't looked back. I had my fair share of issues and there was a learning curve for the administrative stuff but the. For day to day uses Its a rock and couldn't be more intuitive. My girlfriend who is not tech savy thought it was the neatest thing and demanded that I put it on her aging 6 year old laptop that came with Windows ME but had been limping along on Windows XP with a slim 128 megs of ram. Xubuntu loaded even easier on the old lappy and everything worked out of the box including a pcmcia wifi card. It brought new life to a machine that had been used mostly as a coaster for the past couple years. Then came the big upgrade to Hungry Hippo I mean Hardy Heron there were some hiccups but I reminded myself that windows has a similar trouble shooting learning curve that I had 10+ years vested in. The fact that I was able to upgrade to a new version of the OS with such few issues and trouble shoot the ones I had in a couple hours is really a testament to how Robust and friendly Ubuntu and the Ubuntu community has already become. Not only is Ubuntu becoming easier to use but with another few years of experience under my belt I'm sure that fixing the rare problems will be a snap.

Re:Recent Linux convert from Windows

[initdeep]

can you be a recent convert to formatting now? :D

Debian

[name*censored*]

I'd be interested to know why debian was left out - it's widely used, and it's different enough from Ubuntu (despite Ubuntu being a fork of debian).

Poor research

[Spy der Mann]

The guy simply didn't do his homework. For starters, he thinks MEPIS is based on Mandriva - but it's based on Debian. Then, uses the latest beta of Ubuntu to compete with older distros. Finally, there is NO COMPARISON CHART.

What kind of research is that? He just shows a separate review of each distro, to finally announce "and the winner is...". I call this bull. Much more informative is the "girlfriend linux test" article.

Mod article down.

Ubuntu 8.4?

[X.25]

I'd like to order 1 copy of Ubuntu 8.4.

Thank you.

Re:Ubuntu 8.4?

[Culture20]

No problem, it will be shipped in the 40th month of 2008.

Speeds & Feeds Perils

[mpapet]

These kinds of articles harm practically everyone. They eliminate variety and here's how.

1. Focusing on a couple of winners. In Ubuntu's case they've got PR hucksters doing the shouting for them
2. Eliminating new features. These shootouts leave no room for testing new features, programs, etc. It's yay or nay and the nay's always win when something is -really- new.

3. There are a number of "What about distro X, Y or Z?" comments and they are, for the most part legitimate questions. Most of those non-chosen distros simply haven't made a good enough impression in media circles. Those aforementioned "good impressions" usually cost some money.

4. Eliminating new distros. There are -lots- of other linux distros who's first purpose is _not_ a desktop. The problem I'm pointing out is multifaceted and troubling. To boil it down: "Everyone knows that Linux is that other computer system they buy for less and put their stolen XP OS on."

My 2 cents: Debian Testing -still- manages to be completely ignored when it's a good apples-to-apples comparison to whatever new version Ubuntu puts out.

Re:Speeds & Feeds Perils

[discogravy]

as other comments have mentioned w/r/t Debian Testing, it's not a good comparison to Ubuntu; it's central idea is different, which is really what the other replies have been about. Deb Testing is about getting Debian new software and making everything new work well enough that bugs can be squashed. Ubuntu's raison d'etre is about making debian usable for everyday use without making users spend a day looking up config details for their hardware or what chipset their cards are using and what drivers go with what. Testing's cool, but testing's not for desktop users. It can be /used/ for that, but then again, you can also drive cross country on a unicycle, if you're dedicated enough.

Relatively worthless, even harmful, comparison

[MaulerOfEmotards]

This review/comparison is posted May 4th or 5th, when the distros out there are Ubuntu 8.04 release (not beta, and featuring FF 3b5, not b4); Mandriva 2008.1; openSUSE 11.0 beta; and Fedora 9 preview. Thus, the selection of distros compared is outdated already at the time of review, and worse, unfair between distros (bias?). Compounding this, there are factual errors and lack of in-depth coverage.

This review sais very little about the current state of affairs and is of minimal real benefit to anyone not already initad in the Linux world. It might even do a misfavour to newbies wanting to take the plunge.

Admittably, it takes some time testing seven distros on five platforms, but that doesn't change the fact that it fails to represent the actual state of LinuxLand and the distros pitted against each other.

Sadly I've given up

[rAiNsT0rm]

I've been a Linux guy since 1995 and as much as I hate to say it, I have given up. There needs to be a singular distro at the heart of it all which is steered by either Linus or a committee that focuses on one vision and goal. Chaos is great for creating a million cool bits, but not for organizing them into one unified, cohesive unit.

Let's finally get over the aversion to one main distro, or one of each tool and app. No one cares about choice when all it means is 40 buggy half-assed apps and no single solid one. It is a lot of wasted talent, time, and effort. With some direction and drive Linux could surpass anything out there.

Until people begin to wake up, I'll keep it for servers only. Oh, and I'd personally like to thank the genius who decided to go with a beta version of Firefox for a long-term support version of an OS... now THAT is how to FAIL.

And the winner is...

[Brandybuck]

And the winner is... the distro most like Windows!

So 2 versions of RH and no SLACKWARE!?!

[arfonrg]

So, basically, two versions of RedHat were included and no Slackware?

I guess they were scared of Slackware's awesomeness!