How often do we need GNU/Linux releases?

Sam Varghese
Thursday, 02 October 2008 22:35

Opinion and Analysis

When it comes to ceaselessly adding software, which, in a moment of generosity, one might describe as a beta release, then a situation similar to what existed in the Windows world comes about.

What does result is a situation where users have no choice but to wait for the next release to realise the actual potential of a feature that was in the current version.

At times, the feature in question may be removed, because it was introduced in an impulsive moment, in that mad rush to get ahead.

Naming conventions are often used to give the impression that a great deal has happened when, in reality, there have been just minor tweaks.

I remember the time when Patrick Volkerding came out with Slackware 7.0 after he had released Slackware 4.0. Was there that amount of change in the 7.0 version to justify such a quantum leap in version numbers?

No, Volkerding was scared that people would think that his distribution was lagging behind Red Hat which, at that time, was up to version 7.0 or so. These version numbers were often mistakenly used by people to identify distributions and Volkerding was afraid that "Linux 4.0" would be interpreted as being older and outdated compared to "Linux 7.0".

He didn't want people to think that Slackware was in any way behind the development curve.

GNOME, for example, has a six-monthly release cycle. Exactly how much change is there between releases? Is it really necessary to parade such incremental change as a new release and keep trying to induce people to update?

One can understand security fixes being put out periodically, fixes that get incorporated into a release as part of the regular update of any particular distribution.

Unfortunately, when projects follow the correct procedure, they end up getting panned. KDE 4.0 was a major change from the previous version and it came in for a huge amount of criticism simply because a few gentlemen (and I use the term advisedly) did not know that an .0 release is always a development version and not meant for deployment.

It is meant for those who like to live on the bleeding edge. Much in the same way that all the pre-releases of Ubuntu are meant to give an indication of what the final release will be like.

The aforementioned individuals assumed that version numbers mean nothing and that everything happens by small increments. They forgot that, apart from security fixes, it is unnecessary to keep updating software, simply because it detracts from the real reason we use that software - to be productive in some way or the other.

Frequent release cycles do serve one purpose - they give the impression that something is happening. It is similar to a person running in the same spot. A few trivial things can be changed - round windows? - and the appearance of movement is achieved.

This is another similarity to Windows - trying to give the impression that a lot is happening when in reality nothing is taking place.

I think a release once every 18 months is more than enough. Given that interval, there is time to properly test software, there is time to incorporate packages that will really make a difference. Else, it looks like what Indians refer to as naam ke vaste - doing something for the sake of doing it. But I would be interested in what you, gentle reader, think about release cycles.