By Jon Masters
Linux has changed considerably since I first became involved with the Linux community a decade ago. Back then, I had just become the youngest University student in the UK and had gotten access to my first real UNIX box. Shortly thereafter, I discovered I could have my very own UNIX at home for a fraction of the cost of buying my own Sun server kit (which cost around US$ 10,000 new.) A copy of a single floppy disk Linux installation served as enough inspiration for me to spend the next two weeks downloading Slackware over the painfully slow JaNET UK national academic network.
200 floppy disks (and a few duds) later, I had a working installation of Slackware and absolutely no idea how to use or maintain it. It often seemed necessary to re-install the system, as various weird and wonderful error messages would appear every ten minutes or so. And it quickly became difficult to convince friends and family that this was a superior system of the future when the terse output of fsck on a power outage proved enough for another round of re-installation.
After going through one too many such a rinse and repeat cycles, it was refreshing to discover that various local Linux User Groups had gotten up and running. One of the most important early groups in the UK was the LBC - or Linux Buyer's Consortium (later renamed "Linux Emporium," under which the company backing it exists to this day.) This eclectic mix of enthusiasts existed solely in the form of a busy mailing list, on which all manner of topics were discussed on a regular basis. From kernel stability issues to deeply philosophical discussions about the dangers of interfering with electric fences - we had them all on that list.
It wasn't long after that when more regional groups came into existence and started to hold regular meetings. The local Oxford LUG provided a good opportunity to meet others with similar interests, while some of the talks served as an academic basis for further understanding.
Today, there are literally thousands of Linux User Groups (LUGs) around the world. Some towns have more than one, and several major cities have as many as three or four different groups. In the UK, we have some large groups, such as the Greater London LUG (GLLUG), which hold regular organized meetings complete with presentations and even occasional workshops.
Some groups are formed on a purely social basis. (Lonix and Wolves are good examples of social Linux groups.) Other groups aim for a hybrid mixture of business and social interaction. These groups are small when compared to some of the higher profile international groups. Silicon Valley LUG (svlug) probably has more non-valley members than it does actual residents from the local area. The Bombay LUG has 800 (mostly local) members and is still growing quite steadily.
One of the biggest challenges facing LUGs at the moment stems from the fact that a growing number have existed for ten years or more. Many of us know each other quite well, and newcomers often find it difficult to get a word in without feeling that they're somehow offending those to whom their problem seems trivial and not worthy of detailed discussion.
Some LUGs occasionally flare up into completely pointless flame wars that start over trivial macho tendencies of many members to prove their technological prowess at every opportunity. This tendency has lead to some community fragmentation, with a single group forking into a group for veterans and one for newcomers. The problem is that such segregation comes at a cost and is often detrimental to the overall community because it discourages the involvement of newer members who might in time come to pass valuable information to those who come after.
Despite the diverse cultural and geographical distances that separate us from one-another, Linux groups share a surprising number of commonalities - from topics under discussion to wider organizational issues - that seems to be almost universal. It's interesting to observe how many similar conversations occur around the world at any given moment.
This is the first in a regular series of articles on the happenings of the Linux community. Over the coming issues, I'll take a look at the goings on in the various Linux user communities worldwide - from meetings at local LUGs to wider issues affecting the global growth and adoption of Linux and Free/Libre Open Source Software (FLOSS).
This month saw the release of the next incarnation of the Debian-based Ubuntu GNU/Linux, Breezy Badger. Breezy was released according to the six month development cycle that has been used by the Ubuntu project up to this point. The team has decided the next release will be officially supported for up to five years (depending on how it is used - it's three years for desktop users). This timeline introduces a number of potential support headaches. Ubuntu hacker extrodinaire Jeff Waugh proposed a series of measures to mitigate potential problems, including basing much of the next release upon the work that went into Breezy. He even suggested they hold back a kernel release and limit themselves to only updating certain packages. Thus ends the regular resync with Debian's main repository. While GNOME and KDE may well be bumped up a version, it would be unlikely to see radical updates to Samba, Apache, and other server services on the next Ubuntu release.
Following the Breezy release, Ubuntu will hold the latest of its regular six monthly gatherings with a "Below Zero" session in Montreal, Canada. I was present at one such meeting in Mataro, Spain last year, at which a combination of deeply technical, managerial, and social events brought together the growing community of Ubuntu developers. The meeting was also an opportunity to see various drunken escapades and a highly amusing "Linux Show" special, in which a professional Linux entertainer combined various circus acts with informational excerpts.
A big focus of the latest Ubuntu Conference will be the Launchpad sub-project. Launchpad was first announced a year ago as an effort to unite the various bug tracking repositories around the world, providing a single interface through which users and developers can fix a bug once for all the software projects that might be affected by it.
In addition to the technical material, Ubuntu hacker Jeff Waugh has organized an Ubuntu Love Day, at which people can come together to share in their appreciation of all things Ubuntu. These guys enjoy their work a little too much sometimes.
The other major release over the past month was the long-awaited Open Office 2.0. Several Linux distributions - especially Novell's SuSE Linux - have been shipping pre-releases for some time now, so a large number of users will already be familiar with the feature enhancements that accompany the release.
Chief amongst the improvements in OpenOffice 2.0 are enhanced compatibility with Microsoft's Office suite and a long awaited database component that aims to eventually rival Microsoft Access. (See the December 2005 issue of Linux Magazine for a discussion of the Base database application included with OpenOffice 2.0.) Reaction to the release was mostly positive but, on the whole, not as loud as one might have expected, probably due to the fact that pre-release versions are already widely in circulation. Whether OpenOffice's Impress presentation tool will ever fully cope with the mind boggling array of Microsoft Powerpoint presentations remains to be seen, but on the whole, this release is the strongest contender yet.
Meanwhile, the kernel developers have released 2.6.14, and the floodgates have been opened for large numbers of patches to fly in before the stabilization begins on the 2.6.15 kernel.
Under the newly adopted Linux kernel development model, the kernel is developed in clearly defined cycles consisting of a development phase and then several weeks of stabilzation, during which only fixes are accepted. Precisely how many weeks is an issue that has cropped up on several occasions. Recently Linus commented that that the 8 week development for 2.6.14 was at least a week longer than he would have liked. "6 weeks should be the goal-post" according to Linus, "but to hit that, we'd need to shrink the `merge window' from two weeks to just one; otherwise there's not enough time to calm down."
| THE AUTHOR |
|
Jon Masters is a UK-based embedded Linux developer, writer, and consultant. He has been actively involved with the Linux community since starting his first degree at age 13. |
