UpFront

diff -u: What's New in Kernel Development

Zack Brown

Issue #169, May 2008

Linux 0.01 is alive and well. Abdel Benamrouche recently ported the original Linux code from 1991 to GCC version 4. This is the sort of project people do because it's fun and unusual—not because they expect there to be any practical application for it at all. Yet, as often as not, there is. When Cong Wang heard about Abdel's work, his first thought was how useful it might be to university Computer Science departments teaching operating systems. With that thought, he immediately sent Abdel's work to Cong's own CS department. Where they go with it is anyone's guess.

The original Linux 0.01 required GCC 1.40 or thereabouts, according to a linux-kernel post by Linus Torvalds in August 2001, when Tristan Sloughter tried to get 0.01 running on his 386. A little later, in September 2001, Mikulas Patocka actually fixed a bug in the disk request sorting code of Linux 0.01. At the time, Linus offered Mikulas maintainership of the 0.01.xx kernel series, but Mikulas turned it down. Maybe Abdel will decide to take up the banner and maintain 0.01 himself.

The kernel sources include a variety of shell scripts that each try to rely only on the default /bin/sh UNIX shell. Andreas Mohr recently discovered that one of these scripts actually relied on the bash shell, though it claimed to work on whatever the user used by default—that is, on /bin/sh. He ran into this problem when he tried to use the script on a system that used the bash shell by default. So, after doing some cleanup, he submitted a shell to remove all the bashisms from the script. It was not easy—there were a number of obscure bash features represented in the code. But, after some testing, comments from other kernel folks and revised patches, it did seem as though he'd managed to eliminate all the bashisms from the script. Adrian Bunk's suggestion that it might be quicker simply to make the script rely on bash explicitly was ignored in favor of the much more fun project of delving into shell arcana.

Apparently, too many people have started using the new ext4 filesystem. This code is not yet ready for widespread consumption, but it's been in the main kernel source tree for a while already to encourage experimentation. And, folks have been experimenting! Unfortunately, not everyone who's been using it has been aware that it was not fully ready. Adrian Bunk recently reported seeing users trying it out without considering the consequences, just because it was there in the kernel already. To deal with this, he proposed a patch, making ext4 dependent on the BROKEN configuration option. To compile the filesystem, users would have to edit the config files by hand to remove that dependency.

There have been various objections to this, including from folks like Alan Cox, who accused Adrian of meddling too deeply in kernel configuration culture. There is a lot of resistance to making ext4 harder to use, precisely because the ext4 developers very strongly want lots of people to test it. And, as they tend to be kernel “insiders” like Alan, they can get an experimental filesystem into the main kernel tree while other filesystems, who also want lots of testers, have to wait outside the tree and undergo a lot of additional scrutiny before being included.

This is not to begrudge ext4 its place of privilege. The ext4 developers are insiders because they've earned it, and they have a deep understanding of how kernel development should be done. Linus tends to trust their judgment, not because they are insiders, but because they have earned that trust. But, the fact remains that ext4 is in the main kernel tree, and it is not yet ready for regular use. Folks interested in it certainly should test it out if they want to, but with caution (and backups).

Michal Simek may become the official maintainer of the Microblaze kernel port, included in the main kernel source along with the other architectures. He coded up the Microblaze support himself, but he was not very familiar with what would be involved in being a maintainer and what sort of support he could expect from the kernel.org people (such as git repository hosting and so forth). A lot of folks had a lot of advice, and the whole discussion served to summarize current best practices regarding patch submissions and review, and the best way to host a full kernel tree (it turns out that hosting on kernel.org itself has the advantage of sharing git objects with Linus' tree, and this would make for a much smaller repository on disk). It seems likely that Michal will become the official maintainer. There certainly has been enough enthusiasm for him to do so.

A bunch of people have been translating kernel documentation into Chinese, under guidance from Greg Kroah-Hartman, who seems to be leading the effort. Recently, several translations were integrated into the kernel, including some by Li Yang, Zhang Le and Bryan Wu, among others. This push toward greater accessibility has been ongoing for years, but it appears to be picking up speed at the moment. These translations are dramatically increasing the available kernel developers who can participate in Linux development, and they pave the way for a deeper integration with the means of development.

Apache's Share Goes Back above 50%

Doc Searls

Issue #169, May 2008

For more than a decade—from October 1995 to November 2005—Apache's growth in Web server market share went mostly up. In November 2005, the free and open Web server peaked at 70.98% among Top Developers on the Netcraft.com survey for that month.

Since then, the share mostly has gone down. In October 2007, Apache's share declined by 2.8% from the previous month, dropping to 47.73%, while Microsoft IIS gained 2.08% to hold at 37.13%. That was Apache's lowest share advantage since IIS appeared in 1996.

But since then, the trend has reversed again. The latest (February 2008) survey from Netcraft, with January 2008 numbers, had Apache at 50.61% on a 1.04% share increase.

Apache isn't only competing with IIS, of course. Google appeared on Netcraft's survey in 2007 and had a 5.33% share in January 2008.

And, the market isn't a pie. Its size overall constantly grows. The total number of servers, Apaches included, has been sloping upward nearly every month since 1995. One exception is the current (January 2008) report, where Netcraft notes “much slower growth”.

One new open-source server to watch is nginx, or engine x. It's an open-source server developed in Russia. In the Google Online Security Blog in June 2007, nginx had a 4% share (to Apache's 66% and IIS's 23%). The nginx site currently says about 20% of Russian virtual hosts run on its server. On Netcraft, it cruised past 0.5% in January 2008.

He Said It

Doc Searls

Issue #169, May 2008

Instead of the usual They Said It, this month we decided to compile quotations from Linus Torvalds alone, because they show a kind of historic turn as we head into an increasingly mobile Linux-based world.

To the Linux-Kernel Mailing List (LKML), May 25, 2007:

It's Friday evening, and the US is preparing for a long three-day weekend, often considered the official start of summer here.

So what's a pasty white nerd to do? You can't go out on the beach, because the good-looking people will laugh at you and kick sand in your face.

I'm not bitter.

But now you can do something: you can download the latest -rc kernel, and smile smugly to yourself, knowing that you are running the latest and greatest on your machine. And suddenly it doesn't even matter that summer is coming, because you can just sit in the basement, and close the blinds, and bask in the warm light from your LCD, rather than the harsh glare of the daystar.

The geeks with embedded hardware can consider themselves doubly special (and not just because your mothers told you you are), because we've got updates to ARM, SH and Blackfin.

What more could you possibly want?

Source: lkml.org/lkml/2007/5/25/439.

To various IDG publications in Australia in January 2008:

Technology doesn't worry me. Stupid external issues, especially patents and stuff like that—those are the ones that worry technical people. Probably because they feel they can't (including me) do a lot about them. When you don't feel you're in control, you start worrying.

Source: www.zdnet.com.au/news/software/soa/Torvalds-worries-about-patents-and-slow-storage/0,130061733,339285687,00.htm.

The embedded people actually solved a lot of the power problems, but they tended to solve it for their particular platform....You had ten solutions for ten different uses, then none were interchangeable because they were very specialized....Now...we have a good over-arching model that works hopefully for everybody....We're just now at the stage where we can solve them for everybody.

Source: www.zdnet.com.au/news/software/soa/Linux-is-ready-to-go-green-Linus-Torvalds/0,130061733,339285555,00.htm.

I don't even have a mobile phone! I hate phones in general, because I'm the kind of person that when I work I want to concentrate on my work, and if somebody calls me that completely destroys my concentration. I hate phones because they just disturb you, and mobile phones are even worse because you have them with you all the time, so I don't do mobile phones at all. I have one of the early Linux mobile phones in my workroom because I got it for free, but it's not turned on.

Source: www.computerworld.com.au/index.php/id;444282619.

To the Sydney Morning Herald at the same event:

An OS should never have been something that people (in general) really care about: it should be completely invisible and nobody should give a flying f*** about it except the technical people.

It's stupid—when you make a big deal about something like Vista or Leopard, a lot of it is about things I don't consider to be the operating system. It's about the visual shell around it. The fact that Microsoft tied the two together so much actually caused them problems, not just the legal problems. If you manage a thousand clients, or a hundred thousand clients, which is not at all unheard of, you sure as hell don't want to point and click at them. In many ways, Microsoft has had to fix the design mistakes they made when they thought the graphical approach should be a very intimate part of (Windows).

To Microsoft and Apple, the OS is important as a way to control the whole environment, from a marketing and money-making standpoint, to force people to upgrade their applications and hardware.

I don't think they're equally flawed. I think Leopard is a much better system. On the other hand, (I've found) OS X in some ways is actually worse than Windows to program for. That filesystem is complete and utter crap, which is scary. I think OS X is nicer than Windows in many ways, but neither can hold a candle to my own (Linux). It's a race to second place!

Source: www.smh.com.au/news/technology/q-and-a-with-linus-torvalds/2008/02/05/1202090403120.html?page=2.

LJ Index, May 2008

1. Billions of transistors exceeded by Intel's new Tukwila chip: 2

2. Years ago that Intel released a chip with more than 1 million transistors: 2

3. Years ago that Intel released a chip around a half-million transistors: 4

4. Years since Gordon Moore thought up his eponymous law: 33

5. Width in nanometers (nm) of Tukwila's circuitry: 65

6. Maximum read/transfer speed in MB/sec of Intel and Micron's new NAND memory technology: 200

7. Maximum write/transfer speed in MB/sec of Intel and Micron's new NAND memory technology: 100

8. Transfer ceiling of the USB 3.0 spec, in GB/sec: 4.8

9. Position of Russia among all countries searching for “linux” on Google: 1

10. Position of India among all countries searching for “linux” on Google: 2

11. Number of Asian countries in the top ten searching for “linux” on Google: 3

12. Number of European countries in the top ten searching for “linux” on Google: 7

13. Number of North American countries in the top ten searching for “linux” on Google: 0

14. Position of Russian among all languages searching for “linux” on Google: 1

15. Position of English among all languages searching for “linux” on Google: 9

16. Position of Ubuntu among all searches for Linux distros at trends.google.com: 1

17. Position of “Make Ubuntu laptops cheaper than Windows laptops (in all countries)” among popular ideas at Dell's IdeaStorm site: 1

18. Number of entries in Dell IdeaStorm's most popular ideas: 20

19. Number of popular IdeaStorm requests having to do with Linux, a distro or open source: 12

20. Number of device models counted running Linux in Intel's Mobility booth at CES 2008: 5

1–5: ZDNet

6–8: Intel

9–15: www.google.com/trends

16–19: ideastorm.com

20: /photos/linuxjournal at Flickr (models were Clarion, Aigo, Lenovo, Samsung and Digifriends)

OpenID's Open-Source SSO Gains Momentum

Doc Searls

Issue #169, May 2008

The decision by the Identity Gang (now formalized as Identity Commons) to get behind OpenID was an easy one: it was simple single sign-on, or SSO, and already in use at LiveJournal, the popular blogging system created by Brad Fitzpatrick, famous as well for memchached and other fine hacks. Brad also made OpenID open source, making it easy for developers to work with and contribute to it.

One remarkable fact about that Gang meeting (the first Internet Identity Workshop, in 2005) was that leading figures working on other identity systems—people from Microsoft, Sxip, Cordance (i-names) and Higgins—all jumped in to find ways of working with OpenID.

Since then, there have been many workshops, many meetings, much hacking and an acceleration of acceptance toward critical mass. You know that's been achieved when Google, IBM, Microsoft, VeriSign and Yahoo join an organization's board all at once. That happened for the OpenID Foundation in February 2008.

When I asked David Recordon, Vice Chair of the foundation board and OpenID's highest-profile advocate for his take on things, he said, “In 2007 OpenID saw incredible momentum as it grew from a grass-roots technology to a common tool in a developer's arsenal. Already in 2008, it has grown to include support by Google in Blogger and Yahoo by enabling hundreds of millions of their accounts as OpenIDs.”

To find out more, or to get your own OpenID, visit openid.net. See also Reuven M. Lerner's column in this issue on page 18.

Spreadsheeting for Kids and Geeks

Doc Searls

Issue #169, May 2008

The XO is a laptop for children. A product of the noncommercial OLPC (One Laptop Per Child) Project, and run by veterans of MIT's Media Lab, its brainparent is Nicholas Negroponte, who says, “It's an education project, not a laptop project”, and its goal is “to provide children around the world with new opportunities to explore, experiment and express themselves”.

3.5 Geeks per XO Laptop, at the Latest Apachecon

But, I've yet to see an XO in the hands of a child. Nearly all the OLPCs I've seen have belonged to geeks, or have been in use by them. As an example of the latter, see the shot taken at the latest Apachecon.

As it's turning out, XO isn't just a fun toy for geeks and kids, but a target for development as well—for example, the Sweet SocialCalc Project. Writes Dan Brickin:

I purchased an OLPC XO computer during the Give One Get One campaign, which arrived around New Year's. I love my XO and see its great potential. When I tried my new code on it, the code actually ran quite well.

...when we are done I hope we will have native OLPC code driving this (written in Python), so the JavaScript integrates with the OLPC user interface environment (which is called Sugar, hence the word sweet)....

Feedback is welcome, as are volunteers to help us make this project a reality.

In the future, we will also be integrating this code into more traditional platforms for more traditional wiki-like collaboration. Before that, though, I need to complete the implementation of these libraries, adding more commands, functions, etc.

Dan, by the way, is the other half of the pair that created VisiCalc, the first electronic spreadsheet. The other half was Bob Frankston, subject of a feature article this month (see page 42).

For more, visit the Software Garden OLPC page: www.peapodcast.com/sgi/olpc.

LinuxJournal.com

Katherine Druckman

Issue #169, May 2008

We've had a pretty good couple of months over at LinuxJournal.com. James Gray has interviewed interesting folks from organizations such as Lesswatts.org, OSGeo and Mandriva. We appreciate these representatives taking the time to talk with us and share their insights.

Our videos have been quite fun lately, thanks to Shawn Powers. In addition to his usual “gadget” reviews, he has broadened his focus to include reviews such as the open-source game, Battle for Wesnoth. He gave us a quick look at the game and tossed in some bonus footage of himself getting clobbered on screen, so it's definitely worth checking out. If you missed his review of the X-Arcade, that is also worth a look. It will take you back to all those hours spent in arcades in the 1980s. You were there, weren't you? I was! All of our videos can be found at www.linuxjournal.com/video.

As United States politics heat up, we invite you to take a break from the mainstream and join us in supporting an alternative approach this year over at tuxparty.com. There, our favorite mascot will throw out some issues that may not be addressed in conventional politics. We support Tux for president, and hope you will too.

What They're Using: Michael Anti and His Eee PC

Doc Searls

Issue #169, May 2008

Michael Anti is an engineer and journalist whose work has appeared in the New York Times, Huaxia Times, 21st Century World Herald, Washington Post, Southern Metropolis Daily and Far and Wide Journal. He has been a researcher, a columnist, a reporter, a war correspondent in Baghdad (in 2003) and more—and achieved notoriety in 2005 when Microsoft deleted his blog. Today, he is best known for his landmark work on press freedom in China—efforts that have earned him a Wolfson press fellowship at Cambridge University and Nieman Fellowship at Harvard University.

Mike Anti and His Eee PC

It was at a Harvard meeting where I noticed that Michael was using an ASUS Eee PC, with exceptional ease and enthusiasm. Turns out, it's one he bought from Amazon. It came new with Knoppix, but then he “cracked” it to do more than ASUS expects of ordinary users (for example, expanding windows to a full screen). I was impressed by how rapidly he typed on the keyboard. Later I found that he was actually typing in Chinese. I hadn't realized, until he explained it, that it's actually possible to type Chinese at the speed of speech on a qwerty keyboard. “I type in Chinese about five times faster than I write”, he says. The word Harvard, for example, is four keystrokes rather than seven. So, if you know Chinese, you can use it as a kind of shorthand—impressive. (As you see from the photo, he was using Smart Pinyin.)

Typing in Chinese

Using Smart Pinyin

In sum, Michael said he has found the Eee PC ideal for three things: 1) hacking, 2) doing journalistic work and 3) watching TV. (In fact, he believes it is “the future of the TV”.)

Ethan Zuckerman, who was at the same meeting, added, “I've seen these all over the place. I ran into (some) Asian businessmen in Amsterdam last week. And they were all carrying them. It's caught on really, really fast.”

His one caution is adaptation. It took him a week to get used to the smaller-size keyboard. Plus, he adds, “You should have some five minutes to get used to it” when you're coming from a normal-size keyboard. Seems like time he's willing to invest.