Open Ballot: Are distro release cycles too short?

TuxRadar

Back by popular demand, it's our Open Ballot. This is an opportunity to air your views on the important Linux issues of the day, which we'll be chewing over in our regular podcast. We'll read out the most incisive/witty/flamebaity responses on the show, so get posting!

The question is: with many distros adopting a six-monthly release cycle, is this a good or bad thing? Should we be looking at longer development phases so that there are bigger changes between releases, and users don't have to upgrade so often? Or perhaps you think six months is not rapid enough – maybe you'd rather have three months, or abandon releases altogether and just have rolling upgrades. Let us know what you think!

You should follow us on Identi.ca or Twitter


Your comments

No, I like six months

I like to get those performance improvements and extra functionality every six months; the upgrades to my various machines don't cause me much trouble, and people can skip an upgrade if they want.

Spare a thought for the developers, please! Many of them get their biggest kick out of seeing their work being put widely into use, and it would be de-motivating for some if the bus only came along once per year.

6 months is good

6 months is good but I think I am going to have one test machine to see how things are going and two stable machines with it being updated every year.

It depends on your needs

As a previous post stated, it depends on your needs. A few years ago after trying out and having issues with Ubuntu (which is based on debian unstable and testing), I switched back to Debian stable on my computer that I needed to be fast and dependable. I have always run Debian stable on my home server, no problems.

A few years ago I was very "install happy" and loved installing and testing out stuff, now I am happier to have more stability so I can get work done.

With all the distros out there, clearly there should be enough room for the long term stable distros and the new cutting-edge distros.

If you are talking about "on average":
Ubuntu current always seems to have a crazy amount of updates I think for a current release, and should probably have a longer release cycle (but not much longer), while Debian should definitely work to get its stable releases out sooner, which I believe they are doing already.

I think maybe there needs to be more consolidation of distros, that is more of a problem I think, or at least more of an effort to tell me how your new linux distribution Z meets a need that can't be met in distribution X that has been around for awhile and is probably more stable. Clearly the only alternate distro in the past few years that has offered significant improvements over standard distros is Hannah Montana Linux ;).

Anyway, if 6 months has become a standard average for linux releases, then maybe this is the best compromise between stability and cutting-edge features.

I say new versions of

I say new versions of distros should come out once every 6 months and 35 days.

I favor incremental upgrades to existing distros

I favor incremental upgrades to existing distros. The "distro release cycle" approach forces the user (essentially) to throw everthing out & start over re-customizing with each new release.

The release-cycle approach serves those who view tinkering with Linux as an end in itself, but not the great majority of potential Linux users. They see any operating system (Windows, Mac OS, Linux, even DOS!) as a mere tool to get a job done, whether it's word processing at work or Facebook tracking at home. Forcing the average user to switch releases effectively discourages them from using Linux at all.

Also, it appears as if the fixed distro cycle is guaranteed to produce new-distro bugs, which must be fixed after each release. As Linux gets more sophisticated, the number of bugs and interdependencies will necessarily increase over time; while they will get fixed, the short distro-cyclke approach guarantees they will be present.

No, when running a business

Each half year it too often for a Distro. Upgrading Apps would be fine, I work with Apps not with an OS.

The problem is that the developement in Linux is so fast, that one wants to have the newest version while it supports the hardware you desparately need to have support for.

For business it is pretty nerving, one does not have time enough to do upgrades all the time, checking everything etc. etc.

I would prefer a yearly distro-upgrade and more apps-upgrades in between.

I believe so...

I don't like the short release cycles. Its one of the reasons I don't run Ubuntu any more. I quit Ubuntu when Gutsy was released, primarily because the releases were broken enough that users had to completely reinstall instead of upgrade. This is one of the things that drove me away from RedHat was when they broke upgrades around 5.x - 6.x.

I went to Debian back then, and I went *back* to Debian from Ubuntu. I run sid, so I really like the idea of LTS + rolling releases. In actuality, I think the Debian way is really good, with an LTS (stable), a semi-rolling release (testing) with some vetting of packages, and the bleeding edge (sid) for the geeks. I run sid with a very few exceptions (mail server and firewall). On the firewall, I run stable until the packages I use get a little long in the tooth (~6 months), then switch to testing. Since I use the distro names (e.g. etch, lenny, squeeze) in /etc/apt/packages.list, the transition from testing to stable is seamless.

Yes and No

The beauty of open source is that it creates choice. So if you need to have a very stable distro, you can go for debian - long release cycle, not always the most current versions of applications, but very stable. Alternatively, if you want to run the very latest versions of applications, and use all the latest eye-candy, then you can go for a "bleeding-edge" distro, but be prepared to live with a little bit of instability. And there are a multitude of shades of grey (or is it brown, blue and green) in between. The issue of choice is where open source really scores over proprietary software where economics and profit margins dictate that "one size fits all"

2 years, 6 months & Rolling !

I am a great fan of Ubuntu for it is simple to use. I use the 6 month-version but I can understand that some people do want a longer period, so therefor is the 2 year cycle more adequate. For me, I would love to see a third option in Ubuntu, namely a rolling upgrade would make me very happy!
Sure, this could be more unstable, but that would be fine to me, because I long to see the latest builds! Don't mind to be the guinea-pig there!

Entirely too short

I'm a slightly above joe average user. I use my PC to write and produce a small magazine. Web browsing, etc. I love linux and Ubuntu in particular, but every six months is absurd. As someone else state, it's not worth the risk to your data (yes I back it up) and the hastle of doing a full install every six months. Though I do like Ubuntu's model of the LTS version . . . but the last LTS doesn't work on my PC so I had to use 9.04. But that's my point. How the hell can one version, say 7.10 work fine and then 8.04 craps out? Wouldn't boot. That's just frustrating. So, I'll be sticking with 9.04 (okay maybe 9.10 because of the new software install interface and ubuntu one integration - damn!) because it works beautifully.

Having longer periods

Having longer periods between releases means users don't have the hassle of upgrading too often. I would like to see maybe a yearly release pattern but with new software (like new versions of Firefox or OOo) in the repositiries.

You don't rush old ladies in the stairwell.

We, King Mark of the Canonical Free-Flowing $$$ Empire, have decided you need to start it all over again every six (fuckin') months.

And that is good for our Bizness, because it puts Our Name top of the charts every six (fuckin') months.

Why should We care of untested features, unmatured code since We get to decide what is good for Our (fuckin') bizness?

BTW, We also have decided to appoint Us Pope of the FreeSoftware Church, and ye all shall knee down because it is for the Good, and for the Combat against (fuckin') Evil.

We have the (fuckin') $$$ anyway, so We do as We please.

...And now Saint Volkerding just released Slackware13 with KDE4 (YIKES!) and 64bits in 8 , well, FUCKING MONTHS.

How do we stop this disease?
Please?

You don't rush old ladies in the stairwell. Debian, Slack & others: there is no fire in the Desktop. No fire at all in that 0.94% marketshare. Only in Shuttleworth's lower body parts.

Longer Cycle => New Users => More Relevance

Having a desktop user focused release with a longer release & support/update cycle would definitely be a good option in my opinion. there are users who would jump on Linux for their every day desktop if their investment into setting up & fixing issues would pay of longer.

receiving security updates and fixes for 3+ years would be a pro and is also good for "marketing".

why care at all to reach new user groups and broadening choices? reach more people, get more support, eventually more funding and increase to importance of the kernel to big consumer orientated companies (hardware, software, business models...)

just my 5ct ;-)

6 months is far too short

Sny release cycle shorter than 18 months is too short. I would rather have a more stable system than bleeding edge everything.
To me, however, the release cycle is not most important. I think the period of support and updates is equally important to avoid having to upgrade the entire system too often. Although this process is automatic in most distributions, it is a hassle and a risk every time to upgrade ones computers.
So I would prefer to do this rarely rather than often. I think that a support period should be at least 5 years

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Post new comment

CAPTCHA
We can't accept links (unless you obfuscate them). You also need to negotiate the following CAPTCHA...

Username:   Password:
Create Account | About TuxRadar