State of Linux
August 1, 2009 2 Comments
Linux’s world has seen very interesting recent developments. But two that stand out the most for me are Debian’s decision about scheduled freezes and the drama surrounding CentOS. Both can be seen positively or negatively but both, in my view, have improved the state of Linux.
First, let’s look at Debian. I had read Mark Shuttleworth’s interview on Gnome 3.0 some time ago. In it he had mentioned talks between Ubuntu and Debian on freeze dates. Now that Debian has confirmed this idea, it’s official: Ubuntu and Debian can now sit down every two years and discuss what to freeze together. It means that Debian can better plan itself by setting hard deadlines on features. Once that freeze occurs, they can focus on making things as perfect as possible before releasing.
Debian has always been known to release when they are ready to release. It was the Debian way, and worked for their users. If someone disagreed, they could use another distro. This wasn’t very helpful to enterprises as they like to know a semi-formal schedule for maintenance and support before committing to a thing as important as an operating system. All users get good quality but that extra bit – a schedule – especially helps enterprises.
Some would see this move as a betrayal of the Debian-ness of Debian. They would point to the idea of time-based freezes as compromising quality. But they would be mistaken. A freeze is not the same as a release. A freeze means for two years developers work on new features before they spend all their time improving the quality of these features.
Debian previously supported a release for roughly three or four years, not because they set out to do so, but because of their release schedule. Now we know they plan to pursue a two-year cycle and I don’t know how long they plan to support a release. If it’s two to two-point-five years, enterprises might not want to invest in Debian for production systems expected to run longer than that. They would rather than Red Hat’s promise of five to seven years. But if Debian can partially match Red Hat’s promise, it now becomes a very good alternative for enterprises.
This also means that Debian stable and Ubuntu LTS could be very closely related, and users get a choice of which goals they want to support. Do they want the Debian goals or Ubuntu goals? I support both because they both provide choices for different set of use cases. But I also love that I can now use Debian when appropriate and Ubuntu when it fits the bill.
I love Debian because of the sheer amount of choice it offers. Red Hat doesn’t match this choice (because it’s not financially feasible?). I am beginning to like Debian with LXDE as GUI (if required). Red Hat doesn’t have it by default. Yes, there are third parties offering this but if you are paying Red Hat you would want Red Hat to provide these things. They don’t and if you are going to third parties, why not just use something which provides these choices under one umbrella?
Combine Debian with the inroads Ubuntu and Canonical are making, and you have a tasty recipe for success. If we continue to support this alliance with our money, it will continue to become a viable alternative to Red Hat for enterprises. Once enterprises begin to support Debian+Ubuntu+Canonical, hardware vendors are more likely to support this alliance; an all-round victory for all.
I am not against Red Hat. In fact, I am a big fan of their success and contribution. I use RHEL and CentOS whenever I can. But I am also a big fan of using the right tools for the job. And for this idea to work, we need choices.
Now let’s look at the CentOS “debacle”. I have always trusted CentOS to provide the best product they can. Recent events have not dented my confidence. In fact, they have reinforced the idea of not relying on one person for anything. At least two or more people should have access and ability to manage all aspects of a group or project. You know, the “hit by the bus” scenario. CentOS can only improve from this moment.
CentOS provides a great service to all Linux users. It builds on the good work of Red Hat and introduces more users to the benefits of using Red Hat, thus supporting them. All contributors to CentOS should be lauded for their voluntary efforts.
And then there are those who had little confidence in open source to begin with and which has decreased significantly because of recent events. They see open source in general, and Linux in particular, as less reliable than proprietary solutions. I think the reason is not technical but commercial: who will support me if I use this product? I would rather not have such people using Linux then. If they don’t believe in open source, they don’t deserve it either.
I know of the fight between “open source” and “free”. I prefer “open source” because “free”, although a laudable goal, is not practical for all use cases. I would rather everything be “free” but if it’s not workable then I would rather support “open source” than proprietary.
In conclusion, I think Linux is better off today than it was a week ago. Debian is changing itself to keep pace with the changing world. CentOS has gone through a rough period to emerge stronger. Things don’t always remain the same but if they are getting better, it’s a victory for all of us.