The UpFront Section


LJ Index, February 2008

1. Number of x86 processors required to perform the same amount of work as one IBM System z mainframe: 250

2. zSeries mainframe energy consumption as a percentage of that required by 250 x86 processors: 2

3. Percentage of all physical servers that will be virtualized by 2011: 50

4. Number of partners in Google's Open Handset Alliance for its Linux-based Android phone platform: 30

5. Number of Google employees working on the Android effort: 100

6. Millions of mobile phones sold worldwide in Q3 2007: 289

7. Minimum billions of dollars Google will offer in the US 700MHz spectrum auction: 4.6

8. Number of steam engine locomotive makers who succeeded in the diesel engine business: 0

9. Billions of phone lines in the world: 4

10. Billions of mobile phone accounts: 2.68

11. Millions of Bluetooth-enabled device shipments reached in 2007: 800

12. Billions of Nokia phones in use: 900

13. Age of Nokia as a company in years: 142

14. Billions Nokia is spending to become a “consumer Web media company”: 9

15. Billions of mobile phones that will be sold in 2008: 1.3

16. Percentage of 2008 mobile phones that will be sold in Asia/Pacific: 82

17. Linux's percentage of Netcraft's top ten most-reliable hosting companies for September 2007: 50

18. Linux's percentage of Netcraft's top three most-reliable hosting companies for September 2007: 100

19. Linux's percentage of Netcraft's top 48 most-reliable hosting companies for September 2007: 43.75

20. Percentage of the top 48 most-reliable hosting companies for September 2007 that are Linux, FreeBSD, Solaris or F5 Big-IP (BSD-based): 66.7

1, 2: IBM and its Power Estimator Tool, CNN

3: IDC, via Guardian.co.uk

4: The Register

5–7, 12–16: Forbes

8: Bob Frankston

9, 10: Trends in Telecommunication Reform 2007, from the ITU, via Dilanchian.com.au

11: Laptop Magazine

17–20: Netcraft.com

Third-Generation Nokia Tablet Gains a Keyboard and GPS

We can't wait to get our hands on the new Nokia N810, announced last October and released in November 2007 (when we're writing this). Unlike its predecessors, the N770 and the N800, the N810 has one big stand-out (actually, slide-out) difference: a qwerty keyboard. That alone makes it far more desirable, and far more like the departed Sharp Zaurus—a unit some of us still miss.

The other big difference is built-in GPS. The N800, which currently holds the title of Ultimate Linux Handheld (September 2007), required an external GPS receiver connected to the unit by Bluetooth.

On the downside (at least for this writer) is that Nokia has dropped the built-in FM radio featured in the N800. We liked that feature and were looking forward to improvements in it, such as RDS support. Perhaps, if enough of us care, a worthy FM radio will return in a future version.

Meanwhile, we look forward to reviewing the N810 in depth for an upcoming issue of Linux Journal.

Join Us at LinuxJournal.com

We are pleased to invite you to come see what all the excitement is about at LinuxJournal.com. New features and exclusive on-line content help keep our on-line home a pretty lively place.

You can't miss the Web-only content featured at the top of our home page. Here, our authors share the latest information, in-depth tutorials and product reviews. These articles are available only on-line, so be sure to check back often and subscribe to our RSS feed to see what's new.

We plan to bring you new features and relevant stories on-line regularly, and hope you will join us and take part as we continue to build the LinuxJournal.com community.

diff -u: What's New in Kernel Development

Keiichi Kii translated the SubmittingPatches file into Japanese, and Greg Kroah-Hartman passed it along for inclusion in the kernel tree. Greg initially had suggested the translation of this and other kernel docs himself, pointing out that the originals rarely changed, so it should be fairly easy for translators to keep up with them.

Vince Kim submitted a patch to add support for LZO compression to CramFS, using Richard Purdie's LZO kernel library. The result was a performance gain, at a cost of a 10% larger driver binary.

Zhang Wei posted a driver for the Freescale MPC8540 DMA controller, commonly used in routers, switches, printers and similar devices. He also added a corresponding “Freescale DMA Driver” entry to the MAINTAINERS file, listing himself as the official maintainer.

Stephen Hemminger posted a driver, called apanel, to control the panel lights on some Fujitsu LifeBook laptops. He based his work on an earlier effort by Jochen Eisenger, but Stephen's work uses no ioctls or user-space dæmons as Jochen's did. Andrew Morton replied with some minor criticism and praised Stephen's mastery of operator precedence in C. It looks like this driver will be accepted fairly quickly.

Samuel Ortiz posted a driver for the Compaq ASIC3 multifunction chip found in many handheld devices. Andrew Morton ran the checkpatch script on Samuel's patch and found many stylistic problems, which he asked Samuel to take care of and resubmit. Andrew also had technical issues and questions, and Samuel posted a new patch in response.

As he has done many times before, Adrian Bunk made another abortive effort to remove the eepro100 driver that the e100 driver is supposed to replace. He submitted his patch, and Jeff Garzik and Auke Kok pointed out that there were still known problems with e100 that made it not yet a suitable replacement. David Acker, who's been working on these issues, said he would step up his efforts, but he also said that the difficulty of testing specific bugs had generally made the project a lower priority for him at the moment.

Adrian also posted a patch removing three I2C drivers: i2c-ixp2000.c, i2c-ixp4xx.c and scx200_i2c.c; there was no dissent on the list, so this probably is the end of those drivers in Linux.

Adrian also posted a patch to remove legacy I2C RTC drivers that already have replacement drivers in the kernel. But, Jean Delvare said that some platforms still relied on the legacy drivers, and that they should be updated to use the replacements before the old drivers were removed. He alerted the rtc-linux mailing list that the PowerPC platform code should be updated as soon as possible.

Robert P. J. Day posted a patch removing the remaining bits of APUS support from the PowerPC architecture. Some APUS code already had been removed in 2.6.23, and the rest had been listed as broken for more than two years, so it was time to go. No one voiced any opposition to this patch, but Adrian said he also had a similar patch he'd been planning to release soon. It turns out Robert has written a script to find dead code in the kernel, which he runs every once in a while to locate things that can be removed safely.

Gabriel Craciunescu discovered that the TLAN network driver mailing list would accept posts only from subscribers, so he posted a patch to note that in the MAINTAINERS file.

Sam Ravnborg announced the creation of a new linux-kbuild mailing list on the vger servers to replace the old kbuild-devel list on SourceForge and posted a patch updating the MAINTAINERS file to show the new list. The old list could be posted to only by subscribers, and it also was moderated. Sam decided to make the new list after he'd seen too many e-mail messages dropped from the old one. There was not much discussion on linux-kernel about this, but it's doubtful any serious objections will be made. The old list was subscriber-only primarily because of a spam problem that had started when the list first came to SourceForge. Presumably, with the relatively new antispam measures that have been adopted on vger, that problem should be a lot more tolerable.

Having obtained the relevant hardware, Maciej W. Rozycki posted a patch to the MAINTAINERS file, listing the DZ DECStation DZ11 serial driver and listing himself as the official maintainer.

Larry Finger has stepped down from maintaining the b43legacy code and is seeking a new maintainer. He's also offered to give a Linksys WPC54G networking card that has the relevant BCM4306/2 chip to anyone who'll take over the code. Apparently, most of the maintenance requirements involve porting Michael Beusch's b43 patches into b43legacy. Michael also is stepping down from maintaining the bcm43xx code, but a replacement will not be needed, as that code is no longer needed and will be coming out of the kernel at some point.

Searching for Consistencies

We ran across a list of Google search results from June 29, 2004, and thought it might be interesting to compare them with searches on two days in November 2007.

Google Search Comparison

 June 29, 2004November 8, 2007November 13, 2007
Linux 108,000,000234,000,00036,300,000
Free Software7,230,00072,800,00099,000,000
Open Source9,130,000412,000,00074,000,000
Red Hat8,710,00097,900,00012,800,000

Worth noting here is that the November 8, 2007, searches were done in London, but at google.us, to avoid the google.uk site (though tested results were essentially the same when I tried both). The November 13, 2007, searches were done in the US—Boston, to be exact.

Still, the widely varying results make one wonder why the largest deployer of Linux in the world (as well as the world's leading search engine) can't yield more consistent, if not useful, numbers.

Mobile Linux Groundswell

The rumble you hear in the ground is embedded Linux moving into the vast mobile space—filling it not just with more closed devices built on open platforms, but with truly open devices that can connect anybody with anybody or anything, any way they like—and to write and use whatever programs they like, without having to limit usage to the insides of carriers' and equipment-makers' walled gardens.

That's where several harbingers are pointed.

First, there's Nokia's Linux-based N series tablets, now in their third generation with the N810. What matters isn't just that the N line keeps improving (and the Maemo development community right along with it), but that Nokia will sell its billionth phone sometime soon—yet it still needs to cripple most of those devices to fit the customer-containing purposes of its carrier partners' walled gardens. The goal here is to make the Net mobile, and it won't happen until Linux-modeled development methods and values prevail.

Next, there's Linux mobile phone work. In November 2007, Google announced both the Linux-based Android mobile phone platform and the Open Handset Alliance; both add momentum to established open Linux handset development efforts by MontaVista, OpenMoko, Trolltech and others. (Of course, there also are plenty of closed handsets with Linux inside, but those play a lesser role in this movement.)

Next, there's the XO Linux laptop from One Laptop Per Child (OLPC), which finally has begun shipping. The inventive little device is coming in at higher than the originally projected $100 price, but it still has plenty of promise and breaks new technical and cultural ground.

Then, there are efforts, such as the Xandros-based ASUS Eee PC (3EPC) 701 “ultra-mobile” laptop and development platforms from Intel and Via for Mobile Internet Devices (MIDs) and Ultra-Mobile Devices (UMDs), respectively, all of which pave the path toward wanton mobile Linux device development.

Together these suggest that Linux will win in the palm and the ear before it wins in the lap. But, those wins will be bigger anyway, as handheld mobile devices outnumber desktops and laptops by a wide margin. (See this month's LJ Index for some of the latest numbers.)

The gPhone That Isn't: Google Keeps the Linux-Based Handset Market Open

What stood out most in Google's announcement of Android in November 2007 was not that it was a Linux-based open phone platform (there are several of those already), but that it looked, quite literally, like something more than the “iPhone killer” many expected. It was, instead, a closed-phone killer.

Well, not really. The iPhone isn't going to die. Rather, it looked like an open alternative to the iPhone that might be exactly what the phone-makers need to get out of their death dance with carriers and start making cellular telephony (and mobile everything) Net-native.

Apple, for all its inventiveness (which is enormous—credit where due), did a deal with the devil when it launched the iPhone in partnership with AT&T. In so doing, Apple became a captive manufacturer for one carrier and crippled the iPhone's Net nativity. Google, on the other hand, put its enormous market heft behind all phone-makers with the guts to risk breaking ranks with its carrier partners and to start making truly open mobile handsets (a carefully chosen word that means “more than phones”).

The platform is called Android, and the SDK invites development of all kinds of devices, with phones playing the center circle of the market's bull's-eye.

But, the target is much bigger. To explore those dimensions, Google is offering $10 million in awards for developers building mobile apps for the platform.

Challenge I runs from January 2, 2008 through March 3, 2008 (right now, if you're reading this fresh off the newsstand or out of the mailbox). Fifty winners will receive $25,000 toward additional development and will be eligible for ten awards of $275,000 each, plus ten others at $100,000 each. Those are due May 1, 2008, and will be announced at the end of that month. The $5 million Challenge II will reward development on Android-based handsets that will start shipping later in the year. Details for that have not yet been revealed at the time of this writing.

Winners will “leverage all that the Android platform has to offer in order to provide consumers with their most compelling experiences”. If you win (or even if you don't), the intellectual property you create (even if you don't wish to call it that) will be yours to keep.

What's especially cool about the “gPhone” is that it isn't for, or by, Google. That was the defaulted expectation of many, based on expectations set by Apple with the iPhone. Instead, the field remains, as it already was, wide open.


They Said It

Ask not if your company is ready for open source, ask if open source is ready for your company.

—Laurent Lachal, Ovum, said at a conference in the UK, October 31, 2007

Many enterprises are managing quite well with both version 2 and version 3 of the GPL.

—Black Duck Software, said at a conference in the UK, October 31, 2007

The OSS is a meritocracy. If you are the chairman of IBM and you submit a patch to the kernel or to KDE that is rubbish, they will tell you. They don't care who you are, how much experience you have or how nice a guy you are. If you are short, tall, fat, thin, man, woman, OAP or teenager, your code is equally judged on its merit rather than on you. As someone who finds sucking up in business intolerable, this is very refreshing.

I still don't know of a single example of an exclusive platform that worked. Yet companies still try to launch them, ignoring history, and hoping that they can control who gets to make their platform a winner.

The best way to predict the future is to prevent it.

[When] Americans can use the software and handsets of their choice, over open and competitive networks, they win.


White Paper
Linux Management with Red Hat Satellite: Measuring Business Impact and ROI

Linux has become a key foundation for supporting today's rapidly growing IT environments. Linux is being used to deploy business applications and databases, trading on its reputation as a low-cost operating environment. For many IT organizations, Linux is a mainstay for deploying Web servers and has evolved from handling basic file, print, and utility workloads to running mission-critical applications and databases, physically, virtually, and in the cloud. As Linux grows in importance in terms of value to the business, managing Linux environments to high standards of service quality — availability, security, and performance — becomes an essential requirement for business success.

Learn More

Sponsored by Red Hat

White Paper
Private PaaS for the Agile Enterprise

If you already use virtualized infrastructure, you are well on your way to leveraging the power of the cloud. Virtualization offers the promise of limitless resources, but how do you manage that scalability when your DevOps team doesn’t scale? In today’s hypercompetitive markets, fast results can make a difference between leading the pack vs. obsolescence. Organizations need more benefits from cloud computing than just raw resources. They need agility, flexibility, convenience, ROI, and control.

Stackato private Platform-as-a-Service technology from ActiveState extends your private cloud infrastructure by creating a private PaaS to provide on-demand availability, flexibility, control, and ultimately, faster time-to-market for your enterprise.

Learn More

Sponsored by ActiveState