Practical Technology

for practical people.

August 1, 2007
by sjvn01
0 comments

Developers free to work on open-source Atheros Wi-Fi

While Linux desktop users have long been able to use Wi-Fi cards based on Atheros chip sets, they haven’t been able to do so without using proprietary Atheros code. Now, the Software Freedom Law Center has given the green light to developers working on an open-source substitution for the closed-source code.

The technology in question, OpenHAL, enables users with wireless cards based on technology from Atheros Communications to connect to networks using exclusively free and open-source software.

OpenHAL is low-level interface software for Atheros 802.11 Wi-Fi cards. Without OpenHAL, Linux-based systems need a proprietary HAL (Hardware Abstraction Layer) plus a wrapper driver to use these wireless cards. Thus, OpenHAL is an open-source replacement for Atheros’ proprietary HAL.

OpenHAL was initially based on the ar5k driver, which was used as the basis for a proprietary HAL replacement for the OpenBSD operating system. Support for Atheros cards is especially important for desktop Linux, since Atheros wireless chip sets are commonly used by dozens of Wi-Fi vendors such as 3Com, D-Link and Netgear, as well as many laptop manufacturers, such as Hewlett-Packard, Lenovo and Toshiba.

The developers of OpenHAL faced allegations that their project might include material that infringed Atheros’ HAL copyright. The programmers then asked the SFLC (Software Freedom Law Center) to investigate these rumors.

The SFLC agreed to help on a pro-bono basis, and subsequently conducted a confidential audit that carefully compared OpenHAL to Atheros’ proprietary HAL code.

After performing the audit, SFLC concluded that OpenHAL does not infringe copyrights held by Atheros. As a result, OpenHAL development can now continue safely, so long as the OpenHAL developers continue their work in isolation from Atheros’ proprietary code, it said.

“The OpenHAL developers can now continue development with legal clarity,” said Karen Sandler, an SFLC attorney. “We thank Atheros for granting us confidential access to its proprietary HAL source code for purposes of the review. We join Atheros in encouraging developers to avoid proprietary code in their work, using clean room approaches like the techniques used in the development of OpenHAL.”

Historically, Atheros, while not allowing access to its HAL, has been very supportive of open-source development. It is because of this support that MadWifi has long been able to supply Linux kernel device drivers for Wi-Fi cards that use Atheros-based chip sets for 802.11a, 802.11b and 802.11g compatibility. Eventually, 802.11n cards will also be supported.

“Our ultimate goal is to have full support for Atheros devices included in the Linux kernel,” Luis Rodriguez, a Linux wireless developer, said in a statement. “By providing legal clearance, the Software Freedom Law Center has helped us get one step closer to making this a reality.”

“We believe that this outcome will clear the way for eventual acceptance of a new wireless driver into the Linux kernel,” said John Linville, the Linux kernel maintainer for wireless networking.

A version of this story first appeared in DesktopLinux.

July 30, 2007
by sjvn01
0 comments

MEPIS to switch from Ubuntu to Debian

SimplyMEPIS, a very popular desktop Linux, is going to change back to using Debian Linux for its core from Ubuntu. In March of 2006, MEPIS founder Warren Woodford, decided to switch to Ubuntu from Debian for the next version of SimplyMEPIS, version 6.0.

The plan was to use Ubuntu 6.06 LTS (Long Term Service), aka Dapper Drake, as MEPIS’ foundation.

Woodford made this move because Ubuntu has a six-month stable release cycle that will enable MEPIS to offer its customers a dependable release schedule, Woodford explained at the time. “The switch to the Ubuntu pools was made to provide our users with a more stable underlying system,” he said.

Things have changed. “Dapper was not updated in the way our users expected,” Woodford said. “Personally, I think the Ubuntu people spoke sincerely and accurately, but perhaps ambiguously. So there was a misunderstanding among users. The fact is Dapper was updated with security fixes, but not with new versions of the applications.”

At first, Woodford worked around this. “In 6.5 [the next released version of MEPIS] I attempted to create an incremental path forward for MEPIS and Dapper by compiling newer versions of applications for the 6.0 base, and with a lot of work I succeeded for that release. However, along the way I discovered that many of the packages in the Ubuntu Dapper pools were greatly out of date and not very useful to MEPIS and Dapper users.”

You might ask: Why doesn’t Woodford make the six-month releases of Ubuntu, such as the recently launched Feisty Fawn, Ubuntu 7.04, MEPIS’ foundation?

Woodford explained that Ubuntu is rebuilt almost from scratch every six months using source packages from Debian EXPERIMENTAL. “This is a wonderful thing from the point of view of improving the Debian EXPERIMENTAL code. Over time the improvements will make their way through Debian Unstable, Debian Testing and finally to Debian Stable.”

But, Woodford said, for MEPIS users this means that Ubuntu does not provide a foundation for long term incremental upgrades. “Ubuntu is almost a whole new distro each time it’s released,” he said. “By using the EXPERIMENTAL code, each and every time, the Ubuntu code tree is inherently less stable than the Debian code tree, which contains additional levels of testing and vetting and fixing of code.”

Woodford said that sometimes the Ubuntu tree contains desirable newer versions of major user applications, and the major user applications, such as Firefox and OpenOffice, come from upstream sources where the code was already stable.

So it is that Woodford has decided to go back to Debian for MEPIS’ Linux heard. In particular, he will be using Debian GNU/Linux version 4.0 — code-named Etch.

“For the next release of MEPIS, we are using a common core based on Debian. As usual, we will have a MEPIS kernel optimized for performance and out-of-the-box hardware compatibility,” Woodford said.

When it comes to applications, Woodford will be using either up-to-date binaries or compile either the “Debian or Ubuntu variant of the source code, using whichever appears to be the best compromise of up-to-date and stable code.”

Woodford has also found that although Ubuntu is source compatible with Debian, it is becoming less and less binary compatible with Debian. “This was not obvious over a year ago, but it is very obvious now,” he said.

What all this means for users is that, “This new process will allow us to have a major release that can be incrementally updated for two years,” Woodford said.

The first test release of the new Debian-based SimplyMEPIS is expected to appear in early August.

A version of this story first appeared in DesktopLinux.

July 30, 2007
by sjvn01
0 comments

Why Cisco Shouldn’t Drop the Linksys Brand

We’ve known for a while now that Cisco was planning on doing away with its Linksys brand. My question for John Chambers, Cisco’s CEO, is: “Why do you really want to hurt your company so bad?

Back in April, Chambers said, “Over the long run we will move to a Cisco brand in the consumer and commercial space. That will evolve over the next year.” Since then, Chambers reiterated that the Linksys brand was coming to its “end of life” and that in “the evolution of the [brand] it will all come, over time, into a Cisco brand.”

Why? Because, Chambers explained, while the Linksys is well known in the United States, its not well known globally.

And, thats a good reason to kill Linksys, because?

Thanks to my wife being a branding expert, I know a wee bit about branding. Who cares if only a few people globally know about Linksys, if everyone in North America knows the Linksys name?

I asked people shopping in the network section of the Best Buy in Asheville, N.C., what name they thought of when they considered buying network equipment. My totally unscientific survey found that 10 out of 11 said Linksys, with number 11 mentioning D-Link.

You can’t get brand recognition like this overnight. You usually cant get it over years. When people think SMB (small and midsize businesses) or SOHO (Small Office/Home Office) networking equipment, they think Linksys.

It’s really pretty simple. Big networks: Cisco. Little networks: Linksys. Or, as a friend of mine put it, “John Chambers doesnt really think that selling to my dad is the same thing as selling to Rackspace?”

Reading between the lines, I think Cisco is making that mistake. Now, Cisco hasnt officially buried Linksys, so theres still time for them to back up. I really hope they do.

Of course, there’s some overlap between Cisco and Linksys customers. Cisco Smart Business Communications System is an integrated Cisco product engineered just for small businesses. Im sure there are SMB customers for Ciscos Unified Communications 500 Series integrated device, which provides eight Power over Ethernet ports wireless, security (VPN), firewall, etc., at a cost of between $650 to $750 per desktop.

But, you know what? I’m also sure there are smaller SMBs and SOHOs that would prefer to pay less per desktop with a mix and match Linksys solution of say a WRT150N Wireless-N Home Router and a RV016 16-Port VPN Router for less than a hundred bucks a desktop.

Last, but not least, Cisco needs to look around at other technology businesses that thought that SMB and SOHO customers were the same as their enterprise customers. They’re not. They never have been. They never will be.

Technology companies that believe in one-size fits all brand, channel and support customer needs don’t tend to do well. Usually, its the SMBs and SOHOs that get the short end of the stick in these deals. I know Linksys customers aren’t really where Cisco makes the big money. Still, some of those little companies do grow into the ones where Cisco can cash six- and seven-figure contract checks.

So, if I were Cisco, I wouldn’t chance alienating them by killing off Linksys. Theres a lot of life left in the Linksys brand, products and service. And, I, for one, don’t see any down side in letting Linksys live on.

A version of this story first appeared in eWEEK.

July 27, 2007
by sjvn01
0 comments

Open source Thunderbird email client looks for new nest

It’s no secret that Mozilla Corp., the company behind Firefox, Thunderbird, Sunbird, and other open-source Internet programs, has made Firefox its No. 1 priority.

Mozilla CEO Mitchell Baker is now admitting that the popular email client Thunderbird has taken second place, and she’s now looking beyond Mozilla to find another way to advance the program.

In a recent blog, Baker starts with a mea culpa.”Mozilla has been supporting Thunderbird as a product since the beginning of the [Mozilla] Foundation. The result is a good, solid product that provides an open alternative for desktop mail. However, the Thunderbird effort is dwarfed by the enormous energy and community focused on the Web, Firefox and the ecosystem around it. As a result, Mozilla doesn’t focus on Thunderbird as much as we do browsing and Firefox and we don’t expect this to change in the foreseeable future.”

The solution? “We have concluded that we should find a new, separate organizational setting for Thunderbird; one that allows the Thunderbird community to determine its own destiny,” said Baker.

Baker and the Mozilla Corp., however, aren’t sure how they should go about this. “Mozilla is exploring the options for an organization specifically focused on serving Thunderbird users. A separate organization focused on Thunderbird will both be able to move independently and will need to do so to deepen community and user involvement. We’re not yet sure what this organization will look like. We’ve thought about a few different options. I’ve described them below. If you’ve got a different idea please let us know.”

The options are: “Create a new non-profit organization analogous to the Mozilla Foundation — a Thunderbird foundation. If it turns out Thunderbird generates a revenue model from the product as Firefox does, then a Thunderbird foundation could follow the Mozilla Foundation model and create a subsidiary.”

The second option would be to “Create a new subsidiary of the Mozilla Foundation for Thunderbird. This has less overhead, although it still requires a new company that serves the mission of the Mozilla Foundation,” said Baker. Since Mozilla doesn’t have the time or resources to devote to Thunderbird now, it’s not clear how, except in name, this would be different from the current situation.

Option three would be to release Thunderbird as a community project much like SeaMonkey, an all-in-one Internet application that follows in the footsteps of the original Mozilla plans.

In this scenario, Baker sees a small independent services and consulting company being “formed by the Thunderbird developers to continue development and care for Thunderbird users. Many open-source projects use this model; it could be simpler and more effective than a Mozilla Foundation subsidiary. However, creating this as a nonprofit would be extremely difficult. Running a services company as an independent taxable company is the simplest operational answer. We would need to figure out how such a company relates to the Thunderbird product itself. What’s the best way for such a company to release a product? How does that relate to the community project that stays within Mozilla?”

These are all good questions, and Mozilla doesn’t pretend to have the answers. “We don’t know the best answer yet,” said Baker. “And we don’t expect to without a broad public discussion and involvement, which we hope this message will trigger. Today someone suggested to me that perhaps there is another foundation that might be a good home for Thunderbird. I hadn’t thought of this; it’s a creative idea.”

Baker also hints that Mozilla may want to look at a different approach towards email than the traditional PC-based email client represented by Thunderbird. “We would also like to find contributors committed to creating and implementing a new vision of mail. We would like to have a road map that brings wild innovation, increasing richness and fundamental improvements to mail. And equally importantly, we would like to find people with relevant expertise who would join with Mozilla to make something happen.”

“If we can see a path to an innovative mail initiative in addition to supporting existing Thunderbird users, then we are interested in doing so. If we find the best way to improve mail is incremental development of Thunderbird as already planned, then we’ve learned something extremely valuable as well,” said Baker.

Baker then invited comments on the questions she raises either as responses to his blog entry or on the new MailNews: Future of Mail Mozilla Wiki page.

Reactions to these proposals have been mixed. At this point, there seems to be no agreement on which option, if any, should be followed.

Some users are also dismissing the idea of wasting any time or resources on a PC-based email client. Others passionately defend the continuation of Thunderbird. Perhaps the most common theme in the replies has been that Mozilla Corp. is paying too much attention to its bottom line and not enough to the free software goals of the Mozilla Foundation.

A version of this story was published in DesktopLinux.

July 24, 2007
by sjvn01
0 comments

Next major PC company to go Linux will be HP

People used to think the very idea that a major PC vendor would offer desktop Linux was beyond a joke. It was, as Vizzini from The Princess Bride would have said, “Inconceivable!”

But, as events turned out, to quote Inigo Montoya from the same movie, “You keep using that word. I do not think it means what you think it means.”

HP will soon be joining Dell in offering at least one Linux desktop line in its SKU sales listing. Here’s why I believe this.

First, Dell successfully broke the Windows-only wall when it added Ubuntu Linux 7.04 to three systems in its consumer line in May. While Dell hasn’t released any sales numbers, its Linux sales must be doing well. You don’t start offering Ubuntu on another brand-new laptop line and announce that you’ll soon be selling Ubuntu to SMBs (small to midsize businesses) and internationally unless you’re making money from it.

Offering Linux once could have just been a good PR move. Expanding the Linux offering means Dell must actually be selling units.

The other PC companies aren’t idiots. Many of them, like Lenovo, have been toying with desktop Linux for years. Hewlett-Packard has been offering desktop Linux for enterprise customers willing to make special orders via its HP Factory Express service service for anywhere from several hundred to several thousand systems for years.

These customized Linux desktops have been progressively selling better as the years go by. “We are involved in a number of massive deals for Linux desktops, and those are the kinds of things that are indicators of critical mass. So we are really looking at it very hard,” said Doug Small, HP’s worldwide director of open-source and Linux marketing. How big is massive? Try thousands of Linux desktops in a single deal.

HP also already sells any of its Personal Workstation PCs with HP Installer Kit for Linux. With this, users can easily install Red Hat Enterprise Linux. HP also offers SLED (SUSE Linux Enterprise Desktop) on its “xw” series Intel Xeon Pentium 4 and AMD Opteron workstations.

The PC giant has also been slowly moving into offering PCs with preinstalled Linux outside of the United States. For example, in early 2006, HP partnered with Mandriva to preload Linux in 37 countries in Latin America.

HP also already supports and certifies more Linux distributions than any other PC or server OEM. Currently, HP supports: Debian, Mandriva, Novell SUSE, Oracle Linux, Red Hat and Red Flag.

Back in March, Small told me that HP sees “the Linux desktop nearing critical mass,” but the buzz has “not been enough to get a preconfigured Linux desktop or laptop on the price list.”

That was before Dell made its move.

Since then the buzz from people within HP, and I’m not the only one hearing it, is that a regular retail Linux desktop is on its way, and there are other signs that HP is getting ready to take the Linux desktop plunge. For example, my old editor buddy Joe Panettieri reports from Ubuntu Live in Portland, Ore., that Chris Kenyon, Canonical’s director of business development, said the University of Delhi in India is partnering with HP to deploy Ubuntu on as many as 5,000 desktops.

As you might guess from that news, I’ve also been told by sources close to HP and Canonical, the company behind Ubuntu, that you can expect to see a formal announcement that HP has added Ubuntu to its list of supported Linux distributions.

HP also just announced that it would be buying Neoware, a provider of thin-client systems. HP is doing this because it intends to accelerate the growth of HP’s thin-client business by boosting its Linux client software.

Sources within HP partners such as Novell and Ubuntu keep giving me hints that an HP desktop Linux will be on its way sooner than later.

How soon? If I knew that, I’d tell you. But, I think we’re going to see HP announce both business and consumer Linux lines at this August’s LinuxWorld trade show in San Francisco.

What I expect to hear at LinuxWorld is that HP will be offering two Linux desktop SKUs. One will feature Novell’s SLED 10 SP 1 for business users. The other will be for home owners and use Ubuntu 7.04.

If anyone from HP, Novell, Canonical or Ubuntu happens to read this—and I know you do!—drop me a note and let me know if I’m going to get egg on my face or if I’ve called it right.

A version of this story first appeared in DesktopLinux.

July 24, 2007
by sjvn01
0 comments

Launchpad PPA Service: Software Development the Ubuntu Way

During the Ubuntu Live Conference in Portland, Ore., Canonical announced the beta release of its Launchpad PPA (Personal Package Archive) service, a new way for developers to build and publish packages of their code, documentation, artwork, themes and other contributions to free software.

Canonical, the company that backs Ubuntu, uses Launchpad to help develop Ubuntu. The Launchpad program itself is a set of integrated tools that support collaboration and community formation. These include a team management tool, a bug tracker, code hosting, translations, a blueprint tracker, and an answer tracker. Its best feature, the bug-tracker, works by trying to track separate conversations about the same bug in external project bug trackers, such as Bugzilla, Roundup, SourceForge, and the Debian Bug Tracking System.

In this new free offering, individuals and teams can each have a PPA. With this, groups can collaborate on sets of packages, and solo developers can publish their own versions of popular free software. Developers upload packages to a PPA and have it built for multiple architectures against the current version of Ubuntu. Each user gets up to 1GB of Personal Package Archive space, which works as a standard Ubuntu software package repository. Free PPAs are available only for free (“libre”) software packages.

Launchpad itself is not an open-source project at this time. Canonical, however, has recently open-sourced Storm, a generic open-source object relational mapper that is used by Launchpad.

The PPA service is designed to connect developers with their users directly. Users who are interested in those packages can make a single update to their system to enable them to install packages from that PPA. Those users will also receive automatic updates whenever new versions of the packages are built and published in that PPA.

Mark Shuttleworth, Ubuntu’s founder and Canonical’s CEO, explained the significance of Launchpad Personal Package Archives for the Ubuntu community in a statement:

“Many developers want to modify existing packages, or create new packages of their software. The PPA service allows anyone to publish a package without having to ask permission or join the Ubuntu project as a developer. This is a tremendous innovation in the free software community,” he wrote. “We hope that PPA will make it easier for developers and development teams who have excellent ideas to get their work into the hands of users for testing and feedback. They also get to mix with experienced packagers to improve their skills. PPA is a build system, a publishing system and a community experience.”

Matt Zimmerman, CTO of Canonical claimed that PPAs’ also make it easy for developers to test new and experimental software builds.

“Adding a new feature to a package or building it against a new version of a system library requires extensive testing. A PPA allows a developer to form a community of testers who are interested in her changes. The testing community can install the packages, run them for the test period, [and] then remove them cleanly from their system,” said Zimmerman. “If the developer releases an updated version, the Ubuntu Update Manager will automatically notify those testers and enable them to update to the newer versions with a single click. This creates a very efficient environment for developers and testers to improve their favorite software.”

The Launchpad PPA service is currently in beta. To participate in the beta program, would-be open-source programmers should send an email to ppa-beta@launchpad.net.

Launchpad PPA Service will be released for general use on August 22, 2007 in-line with the regular Launchpad release cycle. The PPA Service will be available here.

At launch, software in Personal Archive Packages will be built for x86 and AMD64 architectures.

A version of this story was first published in Desktop Linux.