September 1, 2006
by sjvn01
0 comments

Coming soon: Adobe Flash 9 for Linux

We won’t be seeing it until early 2007, but Adobe is hard at work on bringing Flash Player 9 to Linux. In a series of recent blog postings, Adobe’s lead engineer on the Flash Player team, Mike Melanson, has been keeping the Linux community up to date with Adobe’s plans.

Adobe’s main goal is to create a Flash Player 9 that is feature-comparable to its Windows and MacOS offerings. Another, according to a recent interview with Melanson, is to enable proper audio/video synchronization. In the last version of Flash that was available for Linux, Flash Player 7, this was a significant problem.

Melanson promises that Linux users will get it this time, because, “Adobe is purging … the OSS (Open Sound System) audio API in favor of the ALSA (Advanced Linux Sound Architecture).” While the Adobe engineers have run into some problems with ALSA, they’ve been able to overcome them.

At the same time, though, while Adobe will be using the up-to-date ALSA, it will be sticking with some older technologies that are on their way out. The primary example of this is Flash 9’s support of V4L (Video4Linux) API version 1 for camera input, instead of version 2.

The reason to support V4L1? “Lots and lots of camera drivers provide V4L1 facilities,” and, “comparably few camera drivers provide V4L2 facilities,” Melanson said.

While V4L2 is integrated into the Linux 2.6 kernel, V4L1 is clearly hanging on. As a Video4Linux site states, “A number of drivers don’t support the new v4l2 API yet, so we’ll likely see v4l and v4l2 coexist for some time.”

So why doesn’t Adobe do both? Well, you can only program so many things at one time. Nevertheless, while Adobe isn’t promising anything, it’s possible that the Player will include V4L2 support.

As Melanson said in an earlier blog posting, “To be sure, it’s not that Linux can’t do XYZ, it’s that there are so many ways to do XYZ. And forget about making these different methods interoperate where conducive.”

For the UI (user interface), Adobe is using GTK. But, Melanson said, “People [should] not to get too hung up on this UI issue because Flash keeps its interaction with system UI toolkits to an absolute minimum.”

Still, Adobe is considering native support for both GNOME’s GTK and KDE’s QT. On August 31st, Melanson asked, “Here is a hypothetical: what if the Flash Player were to detect during runtime whether it should use GTK or KDE/QT facilities? The biggest issue here would be reliably and automatically detecting which kit to use.”

Finally, Melanson recently announced that there will be a public beta; it just won’t be as soon as many people would like.

“It will be a beta in the classical software engineering sense — i.e., a version that we believe to be largely bug-free and submitted to the users in the hopes that the last of the bugs will be found and reported,” said Melanson.

“Why are we stubbornly refusing to release, say, an alpha version now” asked Melanson? “Primarily because there are known bugs in the Linux Flash Player, and because we know what the bugs are, and we are on track to fix these known bugs. If we were to release an alpha now, we would likely be inundated with reports about bugs we already know about.”

Look for news of the beta to appear later this fall on Melanson’s blog, Penguin.SWF.

A version of this story first appeared in Linux-Watch.

August 29, 2006
by sjvn01
0 comments

Results from the 2006 Desktop Linux Survey

The results of DesktopLinux.com’s 2006 Desktop Linux Market survey are in, and the votes are all tallied. This first article of a series offers a perspective on how the various desktop Linux distributions fared, and why.

Before jumping into what the survey of almost 15,000 Linux users revealed, though, I should point out a few things.

First, this is in no way, shape, or form a scientific survey. Anyone who wanted to vote could vote. We did make sure that we didn’t have a Chicago-style election with multiple votes from one person — or IP address, anyway — but that was about it.

We also didn’t include some mildly popular choices — Arch Linux, GNOME’s Epiphany Web browser, and the Fluxbox window manager — in the initial version of survey. Our readers spoke early on, and we added them.

Still, when all is said and done, we do think that our survey does say some interesting things about the current state of desktop Linux. So, without further ado, here’s what we found.

The most popular desktop of Linux today is… well, most of you can already guess without seeing the scores: Ubuntu.

Ubuntu, with 29.2 percent of the vote, has been the hottest community Linux since early 2005. While this Linux has had its problems lately, such as the update fiasco on August 21st and 22nd, users continue to download, install, and love it.

And, why not? It’s an excellent distribution. It’s not just users who think this; reviewers have also labeled it the Desktop Linux Champ.

A little closer peek at the data, and some comparison with the Distrowatch page hit list, reveals that “classic” Ubuntu with the GNOME interface is the real winner. Kubuntu, with its KDE desktop, and the educational Edubuntu distributions have their fans, but Ubuntu is what a plurality of Linux desktop users appear to be running today.

In a distant second place, with 12.2 percent, we find Ubuntu’s ancestor, Debian. Close behind it, there’s openSUSE with 10.1 percent of the users. If you included in openSUSE’s totals its corporate big brother, Novell’s SLED (SUSE Linux Enterprise Desktop) numbers, 2.9 percent, the SUSE-twins would be in second place with 13 percent.

After this, we come to what I think of as the first surprise in our survey. Gentoo took fourth place with a total of 9.6 percent. Gentoo, to me, is a Linux expert’s Linux. I know many serious Linux users who work with Gentoo to better understand Linux, but almost no one who uses it as their first choice for day-to-day work.

In fifth place, we find Fedora, Red Hat’s community distribution. Fedora, while still somewhat popular with 7 percent of the vote, seems to have lost some of its charm to users in the last year.

Mandriva, which used to be quite popular, is also no longer as attractive as it once was. Even though we included all its immediate ancestors — Mandrake, Lycoris, and Connectiva — in our count, it still only came in sixth with 4.8 percent of the vote.

Then, there were the others. These are the distributions that did, well, horribly in our survey. GoblinX, a Slackware-based distribution; Pie Box, a Red Hat clone; and Tomahawk, a Linux from Singapore — none of these was able to muster enough votes to even creep up to 0.1 percent of use by our survey partipants.

GoblinX’s poor showing surprised me. It’s well-regarded, attractive, lightweight, and has a small, but active, development community. I really don’t understand why it received so little support. No, it’s certainly not a big name, but still, I’d expected to see it get some support.

Thinking of big names, Red Hat came in with a mere 2.2 percent of the vote. That’s still better than Linspire (1 percent) and Xandros (0.8 percent) did, though.

What’s happening here?

Well, if you take a look at which distributions did well, you’ll see they have one thing in common: they’re all community-based distributions. I think what we’re seeing with our survey is that the people who’ve invested something of themselves in their Linux desktop are the ones voting. The people who simply use the Linux that’s set in front of them, or just buy it, have less invested in it and so are less likely to vote.

I have no doubt, based on all the Linux desktops I’ve seen in use at tradeshows, homes, and offices, that while Ubuntu is number one, SLED and openSUSE are clearly the real number two in the number of users, with Linspire in third place.

I’d be willing to lay down a small bet that as Linspire’s recent community Freespire release gathers more supporters, and based on what I’ve seen of the distro, it will be near the top in our next survey.

I also suspect that one reason why Fedora did so comparatively poorly is that Red Hat recently made it clear that the company, and not the community, is calling Fedora’s shots. If the users and developers don’t feel like they have a real say in what’s going on with a distribution, they’re not as likely to stick around.

Still, I’m sure that Red Hat, in the form of the older Red Hat 9 or RHEL WS (Red Hat Enterprise Linux Workstation) or Red Hat Desktop, is working on many business desktops. As for Xandros, I’ve seen it in too many places to think that it doesn’t have a small, but significant, number of users.

However, without that community boost, I don’t see either one doing well in our open survey. For a realistic look at how many people are using them, we’ll need to wait for the IDGs and Gartners of the world to do a CIO/CTO survey of business Linux desktop use. Come that day, I suspect we’ll find RHEL WS and SLED neck and neck.

For now, though, what I can say with perfect assurance is that Ubuntu is the number one desktop Linux distro, and that the community Linuxes are far more popular with our readers than their commercial brothers.

In my next look at our survey results, I’ll talk about what desktop interfaces and applications are favored by our Linux desktop readers. There won’t be any surprises at who’s on top, but I think you will be surprised by the ones that are closest to catching up with the big names.

A version of this story was first published in 2006.

August 25, 2006
by sjvn01
0 comments

Thank you, Mr. McNealy

Never forget that while he was unable to right Sun in recent years, McNealy wasn’t just an industry giant. He changed the IT world forever. In 1982, Scott McNealy founded Sun Microsystems with three graduate student friends — Andy Bechtolsheim, Bill Joy, and Vinod Khosla from Stanford University.

I doubt they knew they were making history.

Sun’s first workstation was in many ways the world’s first workstation. The Motorola 68000-powered Sun-1 had a network protocol, TCP/IP; a slogan, “the network is the computer”; and an operating system, briefly a port of Version 7 Unix, to be followed quickly by the open-source version 4.1 BSD Unix. This soon became known as SunOS.

The computing world would never be the same.

With that one system, which would launch a billion-dollar-plus enterprise, the foundation network protocol of the Internet was laid. Other companies would also make TCP/IP popular. Sun made it the heart of the Internet.

Workstations, while never as popular as PCs, would for decades be the defining platform of scientists, engineers, and high-end design. When I started working on the Internet in the ’80s, we didn’t use PCs. We, all of us, used workstations, and most of them were made by Sun.

And, while Sun has had its ups and downs with open source, by using BSD Unix it set in motion a culture of bright, inquisitive developers who would eventually turn the software world upside down with open source.

Under McNealy, Sun grew to be a computer hardware giant. Then, when PCs began to erode Sun’s market share, he presided over the transformation of Sun from the workstation company of choice into being the high-end server power.

With the rise of the dot-coms, Sun rose to its zenith.

Always colorful — to put it mildly — McNealy would war with his fellow IT super-CEOs such as Microsoft’s Bill Gates. He was never able to unseat Gates as the top dog of technology, but no one gave it a better, or more spectacular, try.

Unfortunately, while the fall of the dot-coms didn’t destroy Sun, it did almost bring the company to its knees.

McNealy, still energetic, still striving for the top, now ruled over a company that, in its frantic efforts to capture its glory days, kept trying one approach after another — network computers, Linux-powered appliances, and Java.

Some of them — such as the wasted $2 billion purchase of Linux-powered Cobalt Networks in 2000 — only hurt the company. Others, such as the very popular Java programming language, have been technological success stories, but have done relatively little to help Sun’s bottom line.

McNealy was correct when he said, on announcing that he was leaving the role of CEO, that “The time is right. Our product line is fixed … our customers are probably happier with us than they have been in years.”

But it was the stockholders, who watched Sun’s losses mount to more than $4 billion between 2002 and 2005, who were doubtlessly the happiest with the change.

That McNealy would announce that he was leaving Sun on the tail end of a quarter that saw losses of $217 million, or 6 cents a share, compared with a loss of $28 million, or 1 cent a share, in the year-ago quarter, was only too appropriate. It was not the technology that had failed McNealy; it was a technology market that he no longer mastered.

The driving man who had led Sun to the heights in the ’80s and ’90s was not the man who could lead Sun back to the top in the ’00s.

I will miss McNealy. Some may say he won’t really go. That he’ll still pull Sun’s strings as the chairman of the board. I don’t see that. I see him riding off into the sunset. His day, I’m sorry to say, has passed.

But let us not forget, let us never forget, that without Scott McNealy we would have neither the Internet nor the open source that powers so much of it.

Hyperbole? I don’t think so.

I was there in the early days. When the Internet moved from college computer rooms into every home. When open source moved from being an academic curiosity to being a driving engine of software.

As I think of those days, I see Sun workstations and servers — pizza boxes, we called them — running SunOS and Solaris, knitting the Net together. I see programmers tinkering with Unix on SPARCstations and wondering what they could do if only they had the source code. I see, in short, our modern computing world as an infant.

Thank you, Mr. McNealy, thank you.

August 25, 2006
by sjvn01
0 comments

Eric Raymond on desktop Linux

Eric S. Raymond is one of the founders of open-source, and a good deal of Linux’s early popularity came from his non-stop beating of the drum for the free software operating system. Then, a few years ago, he bowed out of the limelight to live his own life.

Recently, however, Raymond shows signs of once more playing a bigger role in open-source circles.

When he appeared at the recent August LinuxWorld Expo in San Francisco, he said several things that some Linux fans will find more than a little controversial. Raymond and I, who go back far too many years together, also talked about the issues in a bit more detail out of the public’s eye.

For example, Raymond believes that within the next two years, the Linux desktop must grab a large share of the desktop market or it will never happen. His logic is that historically users shift operating systems when the hardware platform underneath them fundamentally changes.

That’s what happened when PCs went from CP/M-driven 8-bit computers, like the KayPro and the Osborne, to the MS-DOS 16-bit systems of IBM and Compaq. After that, Windows and Mac OS took over on 32-bit systems from Dell and Apple.

Now, we are moving from the 32-bit world to a 64-bit one. Raymond thinks that if Linux doesn’t grab its share of the desktop now, it will never get the chance.

And, how can the Linux vendors do it? Well, for starters, they must switch over to 64-bit computing as fast as possible.

It will take more than that, though. Linux has been running on 64-bit architectures for years. Indeed, it’s often been the first desktop operating system to appear on 64-bit chips like Intel’s Dual Core Processor and AMD’s Opteron.

Linux, Raymond is sure, also needs, one way or the other, to support proprietary equipment like iPods. “When a twenty-something year old comes up to me and I talk to him about Linux, the first thing he wants to know is: ‘Can I run my iPod on it?'”

Saying, as some do, that Linux should only support non-proprietary, DRM-free (digital-rights-management-free) media formats, like Ogg Vorbis, misses the point. If desktop Linux doesn’t become popular, no one will use it, and the open formats will become mere curiosities.

Thus, Raymond is certain that Linux must, no matter how painful it may be to some open-source purists, support popular proprietary software at least on the level of “user-space” applications. If that means paying licensing fees to the Microsofts of the world so that people can watch Windows media files, then so be it.

Raymond’s position could be described as an open-source version of realpolitik, the political concept that one must deal with issues as they really are, as opposed to how we wish them to be.

So, in the case of iPods, you can object all you like to Apple’s FairPlay DRM, but in the real world, if you want to get people using Linux desktops, you need to enable them to use FairPlay-compliant iPod software.

This is, to say the least, not a popular position with many current Linux users or companies. Of the dozens of popular Linux distirbutions, only Linspire’s Freespire takes this approach. And, this has proven to be a very controversial position.

With Raymond back on the open-source scene, and pushing his own vision of how the Linux desktop should evolve, the arguments are only beginning. Like it or not, the Linux desktop is going to be getting a real kick in the pants.

It will be more than a little interesting to see how this all plays out, especially as Microsoft’s Vista slowly, creakingly makes its way to purported January 2007 launch date.

A version of this story first appeared in DesktopLinux.

August 14, 2006
by sjvn01
0 comments

HP offers Debian support for its servers

HP and Debian Linux may not sound like an obvious pairing, but on August 14 at LinuxWorld in San Francisco, HP announced that it has increased its Linux distribution support options for customers and will now support Debian Linux.

Specifically, HP will be supporting Debian Linux across its HP ProLiant and HP BladeSystem server lines. While Red Hat, Novell SUSE, the Asianux foundation of Miracle, Red Flag, and Haansoft remain HP’s first choice of Linuxes for its customers, it has been working internally with Debian since the Linux community’s beginnings in 1995.

HP will provide Debian technical support for installation and configuration during a server’s warranty period. The company, according to sources close to HP, will also offer “care packs” for Debian in the future. Care packs will essentially be service packs.

This support plan is not wedded to a particular Debian distribution. Debian Sarge will be the first Debian distribution to be supported. Etch, when it’s released, will also be supported.

The company also has no plans to sell, market, or certify Debian on HP hardware. So, why do it?

This new support plan was created to meet customer demand. The new support isn’t just for any server administrator, though — it’s for expert Linux administrators who prefer to use Debian for its cost advantages over commercial Linuxes such as RHEL (Red Hat Enterprise Linux) and SLES (SUSE Linux Enterprise Server). At the same time, though, these administrators do want some help from time to time with specific Debian and HP hardware server problems.

HP will also be shipping, with its fully supported distributions, remote monitoring and diagnosis tools with each HP server.

The company is also going to release a Debian Linux-based, thin client — the HP t5725 Thin Client. This trade paperback-sized thin-client can be customized with thousands of Debian applications to meet the requirements of both mainstream users and specialized industries, such as retail and health care.

“HP is continually working to provide customers with the confidence they need to incorporate and integrate Linux and open source into their enterprise,” said Christine Martino, HP’s Open Source and Linux VP in a statement. “Today’s announcement demonstrates HP’s customer successes and opens the door for even more customers to take advantage of open source.”

A version of this story first appeared in Linux-Watch.

August 4, 2006
by sjvn01
0 comments

Desktop Linux breakthrough: Lenovo preloads SUSE on ThinkPad

Finally. For years, the holy grail of the Linux desktop has been to get a major computer vendor to commit to preloading a Linux desktop. It finally happened.

On August 4th, we found out that Lenovo Group, the company that has taken over IBM’s Personal Computing Division, had made a deal with Novell Inc. to preload SLED 10 (SUSE Linux Enterprise Desktop) on its ThinkPad T60p mobile workstation.

For the first time, a major OEM (original equipment manufacturer) has committed to preloading a Linux desktop.

For years, you could get preloaded Linux from smaller vendors. Linspire, MEPIS, and Xandros all have arrangements with second and third-tier OEMs to produce preloaded Linux PCs and laptops. If you were a big-time customer ordering hundreds of systems at a crack, you could also get a major OEM to preload Linux for you.

Over the years, HP has flirted with offering desktop Linux. For example, in 2004, it released the Compaq nx5000 laptop with SUSE Linux 9.1 as a “test” launch. Dell has slowly moved closer to offering a real Linux desktop, but it’s not there yet.

Of course, both HP and Dell will sell you a computer without an operating system, or with FreeDOS and their best recommendation for a Linux for that system, but that’s really not the same thing.

So, if you were just Joe User in North America, and you wanted to simply order or pick up at a store a brand-name computer with a brand-name Linux, you were out of luck.

That was then. This is now.

The T60p isn’t just any laptop. It boasts a high-end, 2GHz Intel Core Duo processor T2500, with a minimum of 512MB of RAM, which can be pushed up to 2GB of memory. For graphics, it uses an ATI FireGL V5200 with 256MB of RAM.

The system is also expected to come with what Lenovo is calling a “ThinkPad Experience” under Linux, which includes: Access Connections, Configuration Utility, Power Manager, warm and cold docking support for USB and video, and Help Center support.

Ironically, in June, Lenovo was in hot-water with Linux fans because an executive had said that the company would no longer support Linux on its ThinkPad line.

Lenovo quickly distanced itself from these comments. “Lenovo’s Linux strategy has not changed … compared to what the IBM Linux-related strategy, related to the PC environment, was,” said Marc Godin, Lenovo’s VP of marketing for notebooks at Lenovo in an eWEEK interview.

Godin then foreshadowed this day, “We’re about to reinforce that strategy and go beyond what IBM or Lenovo, until now, was doing in terms of its commitments to the Linux community ? and to our business partners who want to use Linux.”

In a way, it’s not a surprise at all. Lenovo has been playing second-fiddle to market-leaders Dell and HP for quite some time. Recently, it’s been making moves to improve its status as a top-tier player by becoming the first to offer an AMD processor-based desktop to the enterprise.

So, adding Linux to Lenovo’s offerings makes perfect sense.

I’ve said it before and I’ll say it again: offering Linux as an OEM — big or small — is a smart move.

First, you automatically get a great deal of interest from a passionate, albeit small, group of desktop users. So, why would you want to bother, since Linux desktop users are only measured in single digits of the market?

Because, when everyone else is offering XP, you need to offer something different. If everyone is selling white bread, don’t you think offering whole wheat or raisin bread might earn you some loyal customers? In this case, loyal customers who will spread the word that you’re a good dealer?

From a user’s viewpoint, this move is great. They can finally just buy a fully-functioning Linux laptop, without worrying whether they need to install this to get the WiFi to work, or to download that to read an Adobe Acrobat document. The system will, or at least it should, be ready to work out of the box.

After all, Linux on the ThinkPad has long been supported first by IBM, and more recently, by Lenovo. In addition, the ThinkPad PC product line has probably had more specific Linux support than any other computer brand. In fact, one of the most useful Linux community sites, ThinkWiki, is devoted to nothing but Linux on the ThinkPad.

While we still don’t have all the details yet, some things I already know.

First, this is a major step forward for the Linux desktop. It may not be as big as when Corel introduced the first user-friendly Linux desktop — now Xandros — in 1999. But it’s big.

Second, I think this is a great move for both Lenovo and Novell. Both companies have been in the doldrums recently. Both needed to shake things up. And, this, this is a move that will shake up not just the Linux desktop market, but the desktop market as a whole.

I think that the Lenovo ThinkPad T60p mobile workstation will only be the first of many more mainstream Linux desktop systems. By itself, it may not look like much, but then so the one fateful snowflake which starts an avalanche.

A version of this story first appeared in Desktop Linux.

Practical Technology

for practical people.