2004 Desktop Linux Summit Presentations

International Expansion of Desktop Linux

A panel consisting of Bernd Kretschmer, Louis Nuages, David Orban, and David van Enckevort discussed the international aspects of desktop Linux. The group discussed various issues relating to the deployment of Linux in the corporate environment. A few generalizations can be drawn from the discussion: The corporate world is not interested in fighting the Linux "crusade". Multimedia capability and the ability to play with installed software may bring down worker productivity significantly. The PC as a platform is increasingly becoming obsolete. Moving off of a Microsoft platform will get more expensive with time.

Here are some pluses and minuses of switching a corporation to Linux:

Areas where Linux has strengths:

• Linux has the potential of a much lower total cost of ownership.
• The Microsoft suit over the Lindows trade name gave Lindows increased name recognition and boosted sales.
• Most standard desktop software is available for Linux.
• Various Windows emulation solutions can deal with some of the harder to port legacy applications.
• Companies with small installation bases can transfer easily to Linux.
• Linux's tighter security is becoming a larger factor for conversion.
• Linux is catching on quickly, the next two years will be very critical for companies considering the switch.

Areas where Linux needs improvement:

• Linux language translations are not always mature enough for practical use.
• The choice of pre-installed systems is small.
• Linux has a pull market, the corporate world is used to a push market.
• Linux has a deficit of available business software.
• Existing systems administrators may only have Windows experience, and may fear losing their Windows skills.
• Market statistics for Linux use are very imprecise, making for difficult decisions.

Componentized Linux

Ian Murdock, the Ian in DebIan, explained some problems with Linux distributions, and gave a brief talk on a concept that he calls Componentized Linux.

Linux distributions have driven the commercialization of the Linux kernel. The distribution has served us well, but it has a few shortcomings. Distributions cause Linux to have some of the same shortcomings as a traditional commercial boxed operating system. Problems from this model include incompatibility issues, system lock-in, the need to create a one-size-fits-all distribution, and a per-seat pricing structure. The consumer is stuck with the vendor's platform evolution decisions and release schedule. Typically, organizations need to supplement a box-install with their own post-install customizations. Dedicating staff time to such work diverts valuable time from the company's primary goals.

Componentized Linux is a concept for automating the assembly of components into a working operating system. The administrator only needs to pick the high-level pieces that the system requires. Using of a set of system construction tools, a customized operating system can be assembled. This allows for an optimized installation to be put together, and creates a base system without a lot of software bloat. These tools are available in two forms, the commercial platform services from Ian's company, Progeny, and a community sponsored platform where more work is performed by the end user.

Ian gave a few examples of installations that can be created with this model; these include server appliances, network attached storage devices, point of sale terminals, and customized desktops.

The word from Doc

Doc Searls from the Linux Journal discussed Neal Stephenson's essay In the Beginning was the Command Line. He talked about the essay's comparison of Windows to a Bigass station wagon, the Mac to a quirky Jaguar sports car, and Linux to a tank (a free one).

He also worked with the idea of Crossing the Chasm, or moving Linux from an operating system that's used by a limited number of specialists to a wider cross-section of the population. Doc kept going with the car comparison, and came to the conclusion that in order to achieve wide-spread use, Linux needs to become the operating system equivalent of the Chevy Cavalier, an inexpensive, generic utility that everybody knows how to use.

Breaking Barriers

Industry analyst Amy Wohl gave an overview of the current state of Linux adoption in the business world and non-commercial settings. She noted that the United States is in a different position than other countries in that there are more desktop systems and more installations of Microsoft software. She reiterated the point that interest and adoption of Linux is high in the EU, Africa, and Asian countries.

Linux adoption is moving forward in the field of education and in the business server area. Adoption in the consumer space is still slow due to a lack of pre-loaded Linux boxes, and the perception that Linux is still difficult to use.

As for the desktop, some of the issues preventing widespread use of Linux include a smaller selection of commercial software, not enough device drivers, and a lack of available training. Microsoft and SCO FUD were mentioned as threats to Linux, she mentioned Martin Taylor, a Microsoft employee who is employed primarily to combat Linux.

Amy analyzed the modern day computer users and usages. Typical corporate users of desktop computers were defined as knowledge workers, clerical workers, and meanderers, people with mobile jobs like sales people and nurses. Desktop computer usage was divided into email, business applications, web browsing, and lastly, productivity applications like word processors and spread sheets.

User reactions to Linux include the purchasing of Linux-based servers, planning for Linux on the desktop, and a bit of skepticism. Some real world examples of Linux adoption were given including wide spread use at IBM, Novell, the European government sector, and US educational institutions.

Barriers to Linux adoption include missing applications, a large base of current Windows users, training for users and administrators, and a sustainable ecosystem. The presentation ended with a rosy prediction that Linux will make it as a mainstream desktop system sometime in the 2004/2005 time frame.

Mainstreaming of Desktop Linux

A question and answer style panel discussion was held on the topic of the Mainstreaming of Desktop Linux. Some of the more interesting highlights are covered here.

Q: What are the priorities for Linux making it into the mainstream?
A: Interoperability with existing standards. Linux developers need to visualize the future desktop. If Linux achieves a great enough market share, Microsoft will have to release its office suite on the platform. Linux needs a more unified user interface standard. More drivers are needed, and more applications, especially in the financial field.

Q: Do the big companies (Apple, Microsoft, Dell) need to participate?
A: The large corporations will follow the users, when there is a demand, they will show up. There is a need for consumers to be able to buy preloaded computers. Customers don't want tools alone, they want support for formats, especially .doc (Microsoft Word). IBM is working hard to create a Linux market. If they move over too late, companies may find that they have no market, an example would be Photoshop and the GIMP.

Q: How will the mainstreaming of Linux happen?
A: By giving it away to schools. Outside of the US, adoption is already high. There is a need for more experts. LUGs can offer free support to local schools. The community should aim to get people using Linux on a daily basis, that will drive home use.

Q: What about support?
A: It is critical to have better support than that offered by Microsoft. Better yet, the distribution vendors should do everything they can to make everything work correctly in the first place. Linux support may be less important. Security was an afterthought in Windows but Unix grew up in the university world where security was always necessary. There are two types of support, user help and bug fixes. Microsoft does poorly with bug fixes, turnaround is glacial at best. Open source security fixes can turn around in 24 hours.

Lastly, the topic of Microsoft XP's more restrictive licensing schemes came up, this may prevent a lot of operating system pirating overseas, the indirect effect may be an increase in Linux usage.

Novell's Enterprise View

Novell's Nat Friedman (of Ximian fame) talked about the evolution of the Linux desktop. He started with a quote from Allexandre Julliard: Even failed companies contribute to the progress of open source.

Nat mentioned the inaccurate mainstream press coverage of Linux, which follows a sine curve, alternating from positive to negative. The motivation of Linux developers was investigated; a study by Eben Moglen revealed that most developers work for intellectual stimulation and learning, not the widely publicized need for "geek status." A lot of Linux developers are motivated by joy, and not by market trends.

The evolution of the Linux desktop was illustrated with a set of slides. Its rapid development was shown from 1992 with basic xterms, to 1995 with a browser and virtual desktop, to 1997 with Gnome, a spreadsheet, and GUI development tools. The present desktop has an office suite, a movie player, and the Ximian Evolution groupware client. Current frontiers are commercial desktop support, and integration of third party applications.

State of Desktop Linux Linspire version

Lindows's CEO Michael Robertson covered the current state of desktop Linux, and filled us in on the current state of Linspire in particular.

Lindows is aiming at the mainstream Linux desktop market; it is directly competing against Windows in the Windows market space. Currently, the company employs around 70 people. Most of the technology problems have been solved, the company has recently shifted from a technology focus to a marketing focus.

The Linspire business model includes:

• Building marketing and distribution channels.
• Selling channel services and proprietary software.
• Selling click-n-run software.
• Selling click-n-buy downloadable music.
• Partnering with 450 system integrators.
• Making money from OEM builders and per-unit install fees.
• Partnering with Seagate to offer bare drives with LindowsOS.
• Making pre-installed systems available in stores.
• Focusing on international markets where Microsoft is not dominant.

The company is in the quiet phase pending their IPO, more details could not be given.

Linspire is based on the Debian distribution and the KDE desktop. The distribution is aimed primarily at desktop users. The company has three main software products, Click-n-Run, Lsongs, and Lphoto.

Click-n-Run is a package management system. It adds a nice GUI layer on top of the basic Debian packages. In addition to installing the packages, it also configures the various desktop icons and menu entries. Software can be installed and configured with one click of the mouse. The software keeps track of the applications that are currently on the system. If newer versions of the installed software are available, they show up in Click-n-Run. A catalog of available software, both free and commercial, is available, along with popularity ratings for the various applications. The package management system aims to make software installation and upgrades as easy as music downloading.

Lsongs is an integrated audio application. The aim is to improve on the current state of Linux audio tools, which currently consist of a large collection of different packages with different user interfaces. Lsongs integrates a music library, CD ripper, CD burner, audio player, streaming radios station player, and music store into one application.

Lphoto is an upcoming (still in beta) image system that is the digital photography equivalent of Lsongs. It is an open-source project. Lphoto includes a digital camera interface, a photo library, a photo viewer and editor. Photographs can be output to web pages, email, and printers.

Just looking at the software, it is easy to see that Linspire is aiming to fill the consumer end of the desktop spectrum. The user experience looks to be pretty polished, if the software works as advertised, it should be a good Linux entry point for a wide variety of new users.

Miscellanea

One demonstration that was quite interesting was Sun's Project Looking Glass, an experimental three dimensional desktop environment. Objects on the desk can be panned and rotated in three dimensions and stacked like books on the side of the screen. Content can be displayed on the front, back, and sides of the desktop objects. The desktop also features floating 3D icons. This new technology may be just barely out of the realm of science fiction, but it looks like something to keep an eye on.

Go back to the 2004 Desktop Linux Summit introductory page.

---

(Log in to post comments)