Which is more important, new features or stability?

[michaeljt]

Fedora is always cutting edge, and needs a lot of picking up of pieces, Debian never breaks but never has new software. I would like to know what the Ubuntu community considers most important - to have the latest features, or that everything works out of the box, or somewhere in-between?

[uberdonkey5]

I don't see computing as an endless progression with more gadgets. People use xp rather than vista because it does what they need and its stable. Jaunty will surely be a step forwards (ext4 etc, boot speed), but with more and more peripherals we need to make sure that everything works well. Stability is a key advantage of linux systems, and we should keep that advantage.

[trixman]

as a new linux ubuntu user i would rather have the os have great stability.

;)

[snowpine]

I think Ubuntu addresses this concern brilliantly by providing a long-term support release, a current release, and a preview of the upcoming release, so that users can pick the version that best suits their needs. :)

[sydbat]

i think ubuntu addresses this concern brilliantly by providing a long-term support release, a current release, and a preview of the upcoming release, so that users can pick the version that best suits their needs. :)+1

[michaeljt]

By the way, since I often hear people talking about the competition between Ubuntu and Fedora to be the first to have new things, a nice quote from Seymour Cray - "One of the problems of being a pioneer is you always make mistakes and I never, never want to be a pioneer. It’s always best to come second when you can look at the mistakes the pioneers made." I would love it if Ubuntu made it a policy never to be first, but sometimes (often?) second.

[overlord.gaurav]

Providing the MOST stable OS will be the best feature, wouldn't it?!
By most stable, I mean, an OS that would run on almost any computer and handle almost any device efficiently!

[ranch hand]

I like something inbetween. As a Blacksmith, I like technology (we invented it). As a long time MS user stability is very tempting.

It is important to get new things out to where real people can use them. It is amazing the things I can screw up.

So if things look good to developers and test good, fling them out there and we will see if I can break them.

[snowpine]

By the way, since I often hear people talking about the competition between Ubuntu and Fedora to be the first to have new things, a nice quote from Seymour Cray - "One of the problems of being a pioneer is you always make mistakes and I never, never want to be a pioneer. It’s always best to come second when you can look at the mistakes the pioneers made." I would love it if Ubuntu made it a policy never to be first, but sometimes (often?) second.

I would have to disagree with your statement that Ubuntu is a "pioneer" distro; in my experience, most of Ubuntu comes straight from Debian Sid (about 6 months after the fact). If Ubuntu was a pioneering distro, it would have OpenOffice 3, for example.

I'm not putting Ubuntu down in any way, just that I see it as refining and stabilizing the innovations of the Debian "bleeding edge" to suit the average desktop user. :)

[Redache]

As a Geek I prefer Bleeding Edge.

As a Human Being I prefer stability.

[Kingsley]

I get both in the latest versions of Fedora.

[uljanow]

I like the approach of the Red Hat people. RHEL has a stable core OS:

kernel-2.6.18
xorg-1.1
gnome-2.16
gcc-4.1
glic-2.5
...

and quite recent userland applications:

latest firefox
openoffice-2.3
openjdk-6
...


But such an approach requires more manpower and isn't well suited for community distributions.

[michaeljt]

I would have to disagree with your statement that Ubuntu is a "pioneer" distro; in my experience, most of Ubuntu comes straight from Debian Sid (about 6 months after the fact). If Ubuntu was a pioneering distro, it would have OpenOffice 3, for example.

I'm not putting Ubuntu down in any way, just that I see it as refining and stabilizing the innovations of the Debian "bleeding edge" to suit the average desktop user. :)

Right, it was just that one of the things that prompted my original post was a discussion on lwn.net about the pressure between Fedora and Ubuntu (probably mostly on Fedora though) to be "first" with new features such as ext4, btrfs etc.

[michaeljt]

The poll results seem to be pretty clear. I would like to extend this with a question to the people who have responded so far (mainly the ones who voted for "in between"). How serious do you think that a bug in a currently supported release should be before fixing it should take priority over working on new features? Like if it causes a system crash, or if it causes an application crash, or data loss, or seriously reduces the usability of a popular application? Those were just examples of course :)

[lukaszr]

stability IMO.

features are nice, but if i keep crashing, what good are features.

[lukjad]

I like sitting in between the two. I like to have new toys, but lets not get so excited to start playing with dangerous toys.

[michaeljt]

The poll results seem to be pretty clear. I would like to extend this with a question to the people who have responded so far (mainly the ones who voted for "in between"). How serious do you think that a bug in a currently supported release should be before fixing it should take priority over working on new features? Like if it causes a system crash, or if it causes an application crash, or data loss, or seriously reduces the usability of a popular application? Those were just examples of course :)

In fact once I have got some suggestions, I may start a new poll based on them, as that may be a better question to ask :)

[snowpine]

The poll results seem to be pretty clear. I would like to extend this with a question to the people who have responded so far (mainly the ones who voted for "in between"). How serious do you think that a bug in a currently supported release should be before fixing it should take priority over working on new features? Like if it causes a system crash, or if it causes an application crash, or data loss, or seriously reduces the usability of a popular application? Those were just examples of course :)

I think Ubuntu should continue to release on a strict 6 month schedule. Bugs can be fixed by updates. Users can read about any bugs that might affect them, and use that information to decide if/when to upgrade to the next release. My opinion. :)

[DeMus]

I would say stability and well tested is more important than having new features. Let it take a little longer before releasing something new, but have it tested, tested and tested again.
The more stable it gets, the better it gets.
Of course Ubuntu needs new features as well as any other O.S. But don't start a race against "competitors" about who is first, start a race about who is the best.
As lukaszr also wrote: what is so nice about new features, when they crash all the time?

Let's go for stability.

[Johnsie]

I would like it if the user had more convenient control of the version of software they use. If Amarok 2 is available then give it to us. Don't make is wait 6 months! I guess that's the price you pay when you let big brother control your software sources.

[snowpine]

I would like it if the user had more convenient control of the version of software they use. If Amarok 2 is available then give it to us. Don't make is wait 6 months! I guess that's the price you pay when you let big brother control your software sources.

Johnsie, the whole idea of a Linux "distribution" such as Ubuntu is that it's a full suite of software that's been bundled and tested to work together. If you are running Intrepid, and I am running Intrepid, then we should both have the same version of Amarok. Ubuntu isn't really an "install only the applications you want starting from a base system" type of Linux.

That being said, you can certainly install the latest Amarok, in several different ways, and you can even do a minimal Ubuntu install and build up from there... it's just not the "use case" for the majority of Ubuntu users. If you are waiting 6 months for something you could easily install from source, blame yourself, not the Ubuntu developers. ;)

[Keyper7]

In my opinion, the third option is the only sensible one. While I do understand the meaning of the other two, they are formulated in a excessively extreme way: new features are not necessarily cutting edge and perfection is impossible.

[jdong]

For me, "in between" doesn't totally capture my needs; What I think is that "it depends on what you use it for":


(1) On my network gateway, if it goes down all 12 computers behind it lost network access. For that system, stability is PARAMOUNT. I don't care if it runs a 6 year old kernel or Apache from the dawn of the century (figuratively speaking), my bottom line is whatever it takes to make sure it behaves in a predictable manner and stays up when I need it to.

(2) On my laptop, I primarily do desktop tasks, and for that I prefer to have the latest and greatest that the FOSS world has to offer, even if that means (A) Things break once in a while (B) I have to spend an hour a week applying and sifting through updates (C) Certain apps crash and require an hour or two of detective work.

(3) On my desktop, I primarily do homework while my laptop is charging or broken. For this system I really do want a balance between dependable stability and a reasonably up-to-date set of applications.

[michaeljt]

I think Ubuntu should continue to release on a strict 6 month schedule. Bugs can be fixed by updates. Users can read about any bugs that might affect them, and use that information to decide if/when to upgrade to the next release. My opinion. :)

Certainly :) But time spent fixing a bug is time which is not spent in advancing the upcoming release, so my question is sort of - how important should a bug in the current release be to merit fixing rather than spending the same time on Jaunty/the current alpha or beta? (Yes, I know a lot of the time spent on Jaunty will also be bug fixing, but there is always the option of leaving a package at the same version as Intrepid.)

[billgoldberg]

I'm on Arch and I got only bleeding edge stuff running.

Everything is stable, nothing crashes.

The only problem I ever had was an update to xorg making X unresponsive. Was fixed by modifying one line of code.

[billgoldberg]

The poll results seem to be pretty clear. I would like to extend this with a question to the people who have responded so far (mainly the ones who voted for "in between"). How serious do you think that a bug in a currently supported release should be before fixing it should take priority over working on new features? Like if it causes a system crash, or if it causes an application crash, or data loss, or seriously reduces the usability of a popular application? Those were just examples of course :)

I voted in between.

If a new version of a non system critical part comes out (lets says open office 3) it should be in the repos after week or two of testing.

New versions of things like xorg should not be implemented.

This seems like a fair deal, but I doubt it will happen.

[Skripka]

I voted in between.

If a new version of a non system critical part comes out (lets says open office 3) it should be in the repos after week or two of testing.

New versions of things like xorg should not be implemented.

This seems like a fair deal, but I doubt it will happen.

Something like the Epic Fail Audio CD Bug on KDE should be patched pronto...but yea, it hasn't even gotten a response from KDE yet. Although SUSE has a patch for it.


If a bug keeps users from being able to use significant functionality of their computer (like mounting and adio CD), they should drop what they are doing and fix.

[adamlau]

Stability, stability, stability. Features are provided for by applications and applications can always be swapped out and replaced as need be. But what good are features if they do not work as advertised, or as expected?

[7mkgw7q]

I don't believe it necessarily has to be one or the other. I don't believe we should sacrifice all levels of stability to be on the cutting edge, but don't think any OS would last very long if they never added new software or features strictly from a standpoint of stability of the system. Both are important.

[Eddie Wilson]

Stability is the most important aspect of an os. Bleeding edge features are not any good if they are not stable in operation.

[svaens]

It is my opinion that features of the operating system (i.e. xorg configuration, audio (pulseaudio), process management, font rendering etc) be as good as possible.
Software could be made available more 'bleeding edge', and users could revert to earlier versions if necessary. But the operating system itself should be rock solid.
While it is cool that we get so much software pre-installed which makes our Ubuntu productive as soon as installed. A distinction should still be made between OS and the productivity tools that may or may not be installed on top.

[chucky chuckaluck]

stability is like getting new underwear for christmas. six months later, you're still wearing it, but you can't remember what else you got.

[jdong]

I think the line is more blurred. If I am doing homework and hit Save and OpenOffice crashes (real bug) or try saving to a different document format and OpenOffice loses both (real bug), or need to use my DVD burner and the burning app crashes on startup in my native language (real bug), do I really care "hey, at least the kernel didn't panic!"? NO! I will call that release of Ubuntu unstable crap and move on to a different distribution or operating system.


And I have done this several times in the past and won't hesitate to do it again in the future if I find my vendor making irresponsible choices in failing to offer a stable well-tested distribution.

[mohitchawla]

I don't really care about new features/applications, but am always tempted to try the latest stable kernel. And anyway, I guess bleeding edge might lure desktop users while stability is what matters for servers. So, both are important in different cases and Ubuntu does just fine in both the cases.

[jdong]

Personally I have no problem working on top of Ubuntu for all my stability needs. I'd be lying if I said I ran Ubuntu.... I think I've replaced a good 20% of the packages installed with stuff I repackaged for my own needs.

Considering that if you like your customized setup better than the next release(s) (and I have at times!), you can keep it for 18 months with security updates, I think it's worth everyone's time and effort to learn how to do simple modifications such as version updates or different-CFLAGS/configure-flags recompiles to existing .debs.

[Mohamedzv2]

I think it's better in between. If it isn't stable enough to use, or it crashes a lot, then I don't want to use it, but at the same time if the software is too outdated or is something that won't fit my needs (and since I'm an artist and a writer my needs are greater than the average user) then I don't want it.

[Blue Beard]

I voted for in between.

To win over people you have to have all the features and toys Microsoft provides to lure them away from their 90+% monopoly.

To get stability you need a large user community and a large development/maintenance group. Commercially it's called get market share.

How many of you have lived through Microsoft glitches? Did you still use Microsoft products after that.

Select the product that suits your needs and get active in the community.

[terabyte1]

Yes, stability is better than innovation - though having said that, I had heard that the ext4 journaling facility was being introduced into Jaunty which would result in faster loading would it not?


Terabyte1:guitar:

[Eisenwinter]

I prefer to have the latest and greatest.

[Rawit]

I love to play and mess with the newest stuff... But lately, for work, I go for stability. Using Ubuntu 8.04 LTS now instead of hopping from release to release.

[cmay]

stable. supported and work out of the box is more important as ubuntu is known to be a beginner freindly distribution and for that great reputation it needs to stay stable and easy to work with. i would simply stop using ubuntu if the easy part and the stability went away for the sake of having all the newer applications. for that i could use crunch bang or any experimenting distribution.

i want my things to work out of the box and for unstable testing distributions i have a pc dedicated for that. for betatesting i been runnning jaunty on it last. it has crunch bang in it right now . in short when i need my stuff to work it should always work. when i want someting that is fun to play with like minix or freedos, jaunty or reactOS i use a dedicated playground for that.

[AlbinoButt]

Reminds me of the old Wii discussion: "What's more important, graphics or gameplay?"

As if the two were mutually exclusive. I want both! I want stability and features. The lack of either is inexcusable in my opinion. Just like "graphics or gameplay". There's no reason why not to have both.

[OrangeCrate]

I voted stability, but life at the edge is fun too...

On my main computer, where I run my consulting practice, I keep two partitions. In one, the current Ubuntu release, just to see what it's all about, and in the other the current LTS release. Remarkably, I end up using the LTS the most. It's the default boot, and I generally don't boot into the current version, unless I think about it. Because it documents the activities of my business life, stability is key on this machine.

But, I do like to live life on the edge too. I recently lost my other computer due to age. On it, I played with other distros, such as OzOS, Arch, etc., and the development versions of Ubuntu. Great fun, and I'm looking for a replacement as we speak.

I guess it all boils down to what you're using the box for. One of mine employs a belt and suspenders approach to stability, while the other lives au naturale on the edge.

[handy]

I voted stability.

I use Arch, which gives both stability & cutting edge. :-)