UpFront

Kids today have it pretty easy. Ten years ago, we needed half a magazine merely to get the basics going. Greg Lehsy's tutorial on configuring XFree86 included a helpful “How to Fry Your Monitor” section, covering things not to do. Configuring wasn't just a matter of editing a tweaky config file—it was a matter of editing a tweaky config file that, if you got it wrong, would destroy your hardware.

We also reviewed two proprietary X server packages with graphical configuration utilities. Both Metro-X and Accelerated-X came up and ran without editing any text.

Dean Oisboid surveyed games for Linux, including the classic Adventure and the BSD games. He ran into a little trouble with Id Software's Doom: “I just wanted to play Doom with sound. Having to recompile the kernel just to get sound had to be insane, but it seemed that this simple yearning for Doom had to develop into a learning experience.” One quick compilation tutorial later, “It worked!”

Advertisers offered everything from full systems to mouse pads and T-shirts. A Burgess Shale of distribution ads included Yggdrasil Plug & Play Linux; Slackware; a Pacific HiTech set with four distributions including Debian; Caldera Network Desktop, which was still based on Red Hat; SoftCraft Linux; LinuxWare from Trans-Ameritech; and S.u.S.E., still with all four periods. PromoX advertised a 100MHz Pentium system with 16MB of RAM and a 540MB hard drive for $2,500.

This 2-D turn-based game offers many strategy choices, no icky blood and guts and lots of challenging maps. You control an army of tanks, helicopters, artillery and other units, and each map gives you a different set of units to command and different victory conditions. The project Web site has a tutorial on creating your own maps too.

The interface is slick and the code is rock solid. Play “hot seat” with two players taking turns, by e-mail or against the computer.

The AI is merciless, even where its own units are concerned. Sometimes you can hold a defensive line against a crowd of computer units and pick off the ones in the middle using artillery. The AI doesn't seem smart enough to fall back to a better position.

Crimson Fields is based on the popular Simple DirectMedia Layer (SDL), a library that you probably already have on your system to support other games.

We're counting on our readers even more this year to participate in the Readers' Choice Awards, and voting will happen in two stages. By the time you read this, “First-Round Voting in 2005 Readers' Choice Awards” (www.linuxjournal.com/article/8266) will be underway on the Linux Journal Web site. The final ballot will be based on the results of that initial ballot. Final voting will take place in July, and the winners will be announced in the November 2005 issue of Linux Journal. As the name says, these are the LJ Readers' Choice Awards, so get on over to the Web site, read the nomination list and send us your votes!

For complete information, details and dates regarding the 2005 Readers' Choice Awards, read “New Procedures for 2005 Readers' Choice Awards” (www.linuxjournal.com/article/8192).

Back in November 2004, Michael Boerner reviewed two of gumstix's tiny little SBCs, built around the Intel XScale PXA255 chip with Linux onboard. Since then, gumstix's product line has expanded, and Michael is following its progress in a two-part review to be featured on the Linux Journal Web site. He tells us, “My last review focused on the waysmall computer (WS200-bt) and its components, as well as using it and its different elements, such as Bluetooth.” This set of reviews focuses on the revised unit, with the addition of Ethernet, the new audio, breakout and external Flash storage modules and actually using the units for embedded applications. In “Testing and Building with the New gumstix SBCs, Part 1” (www.linuxjournal.com/article/8268), Boerner looks at these gumstix components: gumstix connex, a 10-100baseT wired etherstix and a waysmall board, which converts the gumstix unit into a waysmall computer.

“Everybody in Mali uses Linux.” That is no doubt a bit of an exaggeration, but it's a phrase that you'd hear only in a flat world.

—Tom Friedman, in The World is Flat: A Brief History of the 21st Century

A telemedicine centre was installed in Segou and Sikasso, regional capitals of Mali, to improve the ability of these medical centres to both communicate and provide quality medical information. These systems are also being used for “tele-radiology”, where these centres can scan an x-ray, then transmit it to specialists in Europe or the United States for a second opinion.

—Geekcorps Mali, mali.geekcorps.org/article.php3?id_article=52

The capability of computers keeps growing and the number of applications running keeps increasing. The people building the interface keep growing the complexity of that. It's not for lack of effort but the software people are losing ground.

—Gordon Moore, blogs.zdnet.com/BTL/index.php?p=1264

You know, Richard [Stallman] is in many ways a walking advertisement for the advantages of not compromising—when you have a long-term goal whose achievement requires dogged long-term effort over decades.

—Rick Moen, linuxmafia.com/pipermail/conspire/2005-April/000996.html

The promise of open source is to eliminate the choice of products and increase the choice of vendors. Examples: Red Hat, CentOS, Lineox, Tao Linux, Progeny, SUSE, Mandrake, Conectiva, Ubuntu (Canonical), Xandros, Lycoris, Knoppix, Debian. This is the paradox of open source.

—Røml Lefkowitz, from a presentation at the Open Source Business Summit

Linux kernel development is in the process of creating a new push for stability. Greg Kroah-Hartman and Chris Wright have volunteered to maintain a tightly controlled stable tree. The stable trees of the old days would alternate with development trees in cycles measured in years. This new stable tree will exist concurrently with the development tree and will consist solely of important bug fixes. So the 2.6.11 kernel release from Linus Torvalds has been followed by 2.6.11.1, 2.6.11.2 and additional stabilizing releases from Greg and Chris. Even after 2.6.12 is released, the 2.6.11.z tree may continue to stabilize, even as 2.6.12.1 comes out, and so on. Unlike previous stable series, whose maintainers had wide latitude to choose which patches to accept, there are strict rules over what can go into this new stable branch. Even the method of considering patches, and the time between patch submissions and patch acceptance or rejection, is tightly regulated. The stable tree has been dubbed by Linus, the “sucker” tree, because he felt no one in his or her right mind would take on the burden of maintaining it. Chris and Greg have risen to the challenge, and the process is itself still undergoing changes in all aspects. But it does appear that stability is once again a serious target of Linux development.

The SysFS filesystem recently felt a pang of uncertainty. One of the driving forces behind its development always has been to replace the chaos and historical baggage of ProcFS with something clean and sane. The developers hoped that, given a fresh start, the old mistakes could be avoided. Recently, however, kernel folks realized that one of the SysFS directories had been put in the wrong place: /sys/block, it was felt, should really have been /sys/class/block instead. Too late! A great mass of user code already had come to rely on the existing directory location. Greg Kroah-Hartman reluctantly had to admit that the SysFS inconsistency could not be repaired. The first spot of age has appeared on the pristine face of SysFS.

SquashFS continues to try for kernel inclusion and continues to come up short, and its principal developer, Phillip Lougher, grows more and more frustrated. One obstacle appears to be cultural: the kernel developers would like to hear convincing reasons in favor of inclusion; while Phillip has been, as he's said, more concerned with coding the filesystem than selling it. Other obstacles are more technical. Currently, for example, SquashFS has a file size limit of 4GB. Of course, because SquashFS is a compressed filesystem, this really amounts to about 8MB of actual data. Also, it recently was pointed out that readdir() does not return either the ./ or ../ directories for SquashFS, as it does for virtually all standard filesystems. These and other problems continue to thwart efforts to get SquashFS into the standard kernel.

FUSE (Filesystem in USErspace) seems to be on the brink of getting into the main kernel tree, after spending quite a while in Andrew Morton's -mm branch. FUSE has had a rough life, with Linus Torvalds saying for a long time that he thought a user-space filesystem was just inherently a bad idea that never would go anywhere. But, FUSE apparently is turning into the little engine that could and even its detractors are having to step aside. Andrew is turning into a strong proponent and seems ready at last to push it along to Linus.

Several projects have changed hands recently, or their maintainers have become officially recognized for the first time. Pete Zaitcev has been listed as maintainer for both the USB block driver and the Yamaha PCI sound driver. Herbert Xu has replaced James Morris as co-maintainer of the kernel crypto API. And Gerd Knorr has stepped down as the Video4Linux maintainer, leaving that project currently unmaintained.