[LWN Logo]
[LWN.net]

Bringing you the latest news from the Linux World.
Dedicated to keeping Linux users up-to-date, with concise news for all interests


Sections:
 Main page
 Security
 Kernel
 Distributions
 Development
 Commerce
 Linux in the news
 Announcements
 Linux History
 Letters
All in one big page

Other LWN stuff:
 Daily Updates
 Calendar
 Linux Stocks Page
 Book reviews
 Penguin Gallery

 Archives/search
 Use LWN headlines
 Advertise here
 Contact us

Recent features:
- RMS Interview
- 2001 Timeline
- O'Reilly Open Source Conference
- OLS 2001
- Gaël Duval
- Kernel Summit
- Singapore Linux Conference
- djbdns

Here is the permanent site for this page.

See also: last week's LWN.

Leading items and editorials


Succeeding SourceForge. Long-time LWN readers will have, some time ago, grown tired of our pointing out that SourceForge represents a dangerous concentration of free software projects. The site currently claims almost 30,000 separate projects, well over half of all projects in existence. The claim is subject to a certain amount of reality adjustment (how do those 30,000 projects relate to the 500 or so you have installed on your disk?), but the fact remains: a large portion of the free software development community is hosted on SourceForge.

Concern about that concentration of resources appears to be growing. New factors include the continuing financial difficulties at VA Linux Systems and SourceForge's move toward proprietary software (as of this writing, the SourceForge jobs page includes a position for a database administrator to "oversee and deploy the transition from Postgres to Oracle"). SourceForge is an expensive gift from VA Linux to the free software community; if VA continues to bleed cash and continues to move toward proprietary software, the company will eventually be forced to look at ending that gift. No responsible board of directors could do otherwise. The idea of 30,000 projects simultaneously looking for a new home is rather scary.

So the level of concern seems notably higher in recent times. It should not be forgotten, though, that SourceForge has been (and still is) a tremendous act of support for the free software community. The hosting of all those projects has been a major contribution; just as important has been the demonstration of how to satisfy (some of) the community's needs. Through SourceForge, we have learned more about how free software development works, and how to help it to flourish.

So why, exactly, did we end up with a single, monster hosting site? It does not appear that there is a natural monopoly there. SourceForge-hosted projects are essentially independent of each other, and there is little synergy in being on the same server. The simple fact is that there have been few alternatives out there. Almost nobody else has wanted to pay the bills involved with providing that sort of service.

Alternatives are beginning to hit the net, however. The GNU project's Savannah server has been up for about a year, using SourceForge code. Savannah currently hosts 356 projects - smaller than SourceForge, certainly, but significant nonetheless. The Savannah hackers have an ambitious development plan which includes replacing much of the SourceForge code, and taking a new approach to free software project hosting. The new Savannah is drawing some interest, showing up in places like the DotGNU project list.

A crucial part of the Savannah plan is that it does not anticipate creating another huge site to compete with SourceForge. The plan, instead, calls for a distributed, decentralized architecture. Savannah servers would be able to mirror (in a read-only mode) each others' projects, but none would become the One Big Server. A well-defined import/export protocol will make it easy to move projects between servers.

This plan looks like the right one for the future. There is no reason why project hosting needs to be centralized, and many reasons why it should not be. With luck, SourceForge will remain a cornerstone of the free software development community for a long time. But it should not be the entire foundation.

(See also, these other free software development hosting sites: Serveur Libre and Berlios).

GNU-Darwin for the x86. The GNU-Darwin Project has been busy for a while, developing a GNU-based userspace on top of Apple's Darwin kernel (which, in turn, is based on FreeBSD and Mach). The idea, of course, is to build an entirely free system. Much progress has been made in that direction. As of this week, it's possible to try out the results of this project's efforts on an Intel-based system: the GNU-Darwin x86 port is now available. Interested users can do an installation over the net, or from a CD image.

Once upon a time, running any sort of Unix-like system on consumer-level hardware was difficult and expensive, when it was possible at all. Now, instead, the situation is rather different. Users can choose between the BSD variants, the GNU HURD (someday), GNU-Darwin, and, of course, an unbelievable number of Linux distributions.

There does appear to be a solid level of interest in free systems which are not based on the Linux kernel. One might well wonder why people would go to the trouble of building a new, GNU-based system on a new kernel, when Linux works so well. What's the payoff?

One benefit, certainly, is the joy of working with new and cool software. Not everybody likes the design decisions built into the Linux kernel, and many of those people are attracted to Mach-based systems. Linux runs on much Apple hardware, but users of that hardware will certainly see some appeal in running a kernel supported by the vendor. And, of course, it's a fun toy.

Then, some users have other reasons for wanting a free system without the Linux kernel. The Free Software Foundation has long felt that Linux has stolen much of its credit. The battle to rename it GNU/Linux has not gone all that far, and resentment remains. The same spirit that causes FSF developers to push forward with HURD development also draws their attention to other, non-Linux alternatives.

The interesting thing is that, at the user level, the tools are the same. Very few users have an attachment to the Linux kernel itself; they want a free system that reliably does what they need. Perhaps, in the future, the vision of a system called "GNU" will be realized, with multiple kernels provided as installation options.

This LWN.net weekly edition comes out one day early so that the LWN staff may enjoy the (U.S.) Thanksgiving holiday. We'll return to our normal publishing schedule next week.

Linux Kongress 2001 will be held November 28-30 in Enschede, The Netherlands. LWN editor Jonathan Corbet will be present and speaking on 2.5 kernel development. A good time should be had by all; see the Linux Kongress web page for details on the event.

Inside this LWN.net weekly edition:

  • Security: Quickstart security guides; OpenSSH 3.0.1.
  • Kernel: Synchronous file operations; devlinks
  • Distributions: Whither Progeny?; SuSE Linux 7.3 is news.
  • Development: Psyco, the Python Specializing Compiler, Ganymede 1.0.8, OpenNMS 0.9.0, Evolution 1.0 rc2, Gnumeric 0.76, Python 2.2b2, Jext 3.0.
  • Commerce: Alan Cox a candidate for Technology Review 100; The resurrection of OpenMail.
  • History: UserFriendly launched; LWN at Comdex.
  • Letters: Kernel development; dealing with bugs.
...plus the usual array of reports, updates, and announcements.

This Week's LWN was brought to you by:


November 22, 2001

 

Next: Security

 
Eklektix, Inc. Linux powered! Copyright © 2001 Eklektix, Inc., all rights reserved
Linux ® is a registered trademark of Linus Torvalds