Desktop

Digg This!

Public and commercial Linux distributions already contain most of the functionality expected by most users from their computers. However, when it comes to Linux establishing a beachhead on the desktop, the battle turns to applications, applications, and more applications.

Certainly Linux distributions contain hundreds of useful desktop applications, but the lack of familiar productivity apps, the problems with document interchange between system platforms, the delayed support of popular desktop hardware, and the lack of preinstalled product offerings have limited the growth of Linux on the desktop to task-specific markets. In response, the Linux Desktop Community is now coming together to address the problems of application availability, desktop integration, and access to new desktop hardware and mobile devices.

The Linux Desktop Community is comprised of a plethora of desktop organizations, known as the desktop dot orgs, Linux distributors, applications/middleware vendors (ISVs), hardware/platform vendors, and standards organizations. The OSDL Desktop Linux (DTL) working group recently hosted a group of desktop architects from these organizations where significant progress was made towards collaborating to solve common problems in the following desktop Linux areas: (1) independent software vendor (ISV) and developer support and (2) Open Source (and mainlined) drivers. Addressing these common focus areas accelerates Linux in the desktop markets and the benefits extend all of the Linux markets (i.e., servers and mobile).

The organizations in the Linux Desktop Community have developed statements describing how to keep the dynamics of the Open Source development process while working together on common goals. ISVs producing both Open Source and proprietary software are looking for greater clarity and direction. Users are looking for better hardware support and advanced graphics capabilities. System integrators are looking for easier ways to roll out and support Open Source desktops. To reach these goals it's increasingly necessary for the Linux Desktop Community to work together as a stronger team. The industry needs a unified approach to the hardware driver challenge. We need to pool our relatively scarce graphics expertise to extend the relevant systems we share to the next level. We also need to agree on which common, non-differentiating technologies to share to increase consistency without diminishing our individual projects' identities and goals.

Linux Desktop Community statements:

• A common stance:
  http://groups.osdl.org/workgroups/dtl/desktop_architects/public_statement
• A Commitment to Commonalities:
  http://groups.osdl.org/workgroups/dtl/desktop_architects/community_statement

Significant energy, commitment, and desire in the Linux Desktop Community to work together have the potential for significant change in the Linux desktop landscape in 2006. The momentum in the focus areas identified by the desktop architects will be described in the remainder of this article.

Desktop Architects at December Meeting
ISVs and Developers
End users don't care deeply about which operating system they're running; they just want and need to be productive. They simply care that their application or set of applications run on their desktop and that they can access their investment of data generated through these applications. Thousands of Open Source and proprietary applications exist which could be used on a Linux desktop. However, once users are accustomed to their application(s), they are hesitant to switch to an Open Source equivalent or to an Open Source platform that might not support their applications.

So what can be done to stimulate developers and ISVs to develop or port applications to Linux? A variety of market factors present a chicken-egg dilemma. Application vendors may not port their application to Linux until there is sufficient market share to justify the investment, and there won't be sufficient market share until key applications are ported to Linux. Desktop architects realize that they can't themselves address the market issues and instead choose to focus on easing application development and porting to the Linux desktop. Perhaps the biggest impediment for application developers is that they have to decide which desktop environment(s) to support. In some cases, they may only be able to justify a single desktop environment such as KDE or Gnome. If a single port of the application would work on any or all of the desktop environments, it would be a huge gain for the application vendors and keep porting costs down.

The Portland Project was born at the Desktop Architects Meeting in December 2005 to generate a common set of desktop interfaces and tools so applications can integrate across desktop environments. The approach is to create programming interfaces (i.e., libraries and tools) that applications developers will use to access desktop capabilities and assets. This programming interface will abstract the application from the desktop environment specifics. Software vendors won't have to generate different application packages for different desktop environments.

The Portland Project is working in concert with the Linux Standards Base (LSB) to document and standardize the application programming interfaces. The Portland Project interfaces won't break any existing applications interfaces supported by desktop environments. In other words, existing applications will continue to work even if they don't use the new Portland Project interfaces.

When will application vendors get to start using these interfaces? The infrastructure will emerge first (implementation of the components in the model above). The development team is experimenting with prototypes now. Some courageous application vendors are testing these prototypes. Jeremy White, a developer at Codeweavers, recently commented, "I'm cheerful to volunteer as an ISV guinea pig." Lubos Lunak, a key Portland Project developer responded immediately by pointing Jeremy to the code, telling him about the handful of API calls that have been prototyped, and asking for suggestions and feedback on the interfaces. ISV feedback is critical for the developers as the Portland Project interfaces mature from prototypes to standardized interfaces.

Source code: http://webcvs.freedesktop.org/portland/portland/
Mailing list: http://groups.osdl.org/workgroups/dtl/desktop_architects/mailing_lists/
Project Page: http://freedesktop.org/wiki/Portland

The first capabilities to emerge from the Portland Project will likely be things like the consistent installation, removal, and updates of menu items, installing an application launcher to the desktop, screensaver management, and associating applications with specified MIME-types.

Open Source Drivers
Linux always seems to lag the rest of the market in supporting new hardware. Most hardware vendors write new device drivers immediately for Windows, but drivers are often available for Linux systems only after the Linux community writes them. This lack of new hardware support is amplified on the desktop since new desktop and mobile devices are now emerging daily. One of the biggest hurdles for device driver developers entering the Linux Open Source community is acclimating themselves to a different development environment and learning how to engage with the existing community. These developers need information to help themselves get accustomed to the methods, processes, and coding styles of the Open Source community. OSDL, in concert with the OSDL working group initiatives has established an Open Source Device Drivers site. The focus of this site will be to act as an information portal to developers, testers, and users to help them gain the vital knowledge they need to begin their interaction with the Open Source development community.

Open Source Device Driver site: http://developer.osdl.org/dev/opendrivers/
Mailing lists: http://developer.osdl.org/dev/opendrivers/wiki/index.php/Mailing_Lists

The Linux Desktop Community is also coming together around desktop printing, wireless capabilities, and power management. These communities are developing a common vision across the desktop organizations and are meeting in mini-summits.

Mini-summits: http://groups.osdl.org/workgroups/dtl/desktop_architects/

Till Kamppeter leads the Desktop Printing group. Till manages linuxprinting.org, which provides resources to help with printing under free operating systems such as Linux. Printing on Linux should just work. The desktop printing community is addressing installation issues, usability issues, and implementation issues such as which job transfer format to use across systems. These issues are being solved across Linux distributions and across desktop environments.

Steve Hemminger, a networking specialist at OSDL, leads the Wireless group. Wireless networking support has been a persistent challenge. The problem is well known in the kernel community but difficult to solve. The wireless developers are addressing the following topics:

• User interface - management and configuration
• Wireless security
• Kernel architectural issues

Significant problems are being addressed in understanding the regulatory issues and developing a unified architecture for wireless networking. John Linville has recently agreed to be the wireless networking maintainer for the Linux kernel. Several different wireless driver architectures each only support a couple of different wireless chipsets. Since John took over, integrating these separate protocol stacks into a unified solution has been a top priority.

Pat Mochel leads the Power Management group. Things are heating up in power management because hardware is getting more diverse, devices are getting smarter, and users care about it. Power management developers are addressing the following topics:

• Embedded - making Linux run better in constrained spaces
• Runtime - improving the overall power consumption efficiency of Linux
• Policy and interface - making it easy for the user to select a power management policy and exploiting the hardware for efficient power consumption
• Infrastructure - driver and platform infrastructure to make it all work together
• Suspend/resume - making it work

OSDL's Desktop Linux Initiative
The OSDL Desktop Linux (DTL) Initiative was formed when OSDL members recognized the need to address the desktop infrastructure, especially for the enterprise markets. The goal of the OSDL DTL has been to work to accelerate the development and adoption of Linux on the desktop. Linux is being deployed most successfully in task-specific environments where there is little reliance on legacy and office productivity applications. The use of Linux has been growing in fixed-function products (ATMs, airline boarding, kiosks, point-of-sale terminals), transactional workstations (travel agents, call centers, bank administration, government forms), technical workstations (CAD/CAM, movie animation), and basic office users (document review, simple office productivity). OSDL's DTL working group efforts and the efforts of folks community-wide will certainly increase this growth and integration. Collaboration is key.