End-user computing on the desktop is perhaps the most eagerly examined part of the open source stack. Thousands of megabytes of Internet bandwidth have been spent discussing the relative merits of the various options available, along with when and if they will be ready to replace the domination of Microsoft Windows and its Office suite. The good news is that they are. The bad news is that, as with other solutions in the open source space, caveats apply.
A large body of software exists in this space. Here, we will focus on the major areas that are of immediate use to the enterprise. In some cases, multiple open source alternatives to proprietary desktop applications are available. We evaluate and recommend alternatives based on attributes of value to an enterprise: ease of use, consistency, a well-defined migration path from vendor alternatives, and continuing interoperability with popular Office document formats.
We also focus on five primary user segments:
Users who run only one application on the desktop and use the desktop only for that application.
Users who run industry-specific desktop applications, and don’t care about the operating system or windowing environment being used to run the application. They might use the desktop for basic email and web browsing as well as for instant messaging.
Users who run several structured or forms-based web applications as their primary interaction with the desktop. They also use email and a web browser. Once again, the choice of operating system and windowing environment is secondary to the user experience.
Users who rely on the desktop to implement business processes and for productivity applications (primarily word processors, spreadsheets, and presentation software). These are not power users, but they do collaborate extensively with users in other departments of the organization and with other organizations that are using interoperable document formats (the Microsoft Office suite, for example).
Users who depend on the desktop for various Windows applications and are tied strongly to Microsoft Office formats.
Solutions for the desktop space work in layers, with solutions at each layer depending on functionality and services from the layer below and providing functionality to the layers above. The key layers of the desktop are shown in Figure B-1.
We see two main migration scenarios:
In this scenario, we see a Linux distribution and a desktop environment and the productivity applications being replaced simultaneously. This is best accomplished by first doing a pilot among a representative subset of the user base.
This scenario involves replacing the browser and Office application suite first, and then migrating the users to a different desktop environment and operating system.
A variety of Linux distribution choices are available for enterprise-wide deployment. The more cutting-edge distributions provide a lower level of support. But several mainstream options also are available, including Red Hat, Mandriva, SuSE, Debian, and Gentoo. The first three are company-supported distributions that come with the option of purchasing support contracts. The last two are community supported, mostly through mailing lists, forums, and chat rooms. When choosing a distribution, you should apply the following criteria:
All Linux distributions provide a method of distributing upgrades and updates. Red Hat, for example, provides updates via RPM. This is the single most important criterion when choosing a distribution, as it directly affects how easy (or hard) it will be to maintain the deployment base.
Some distributions are more polished than others. Depending on the technical skill level of the user base, a more polished distribution might be desirable.
Some Linux distributions are available with company-backed support, and others offer only community-backed support. The system administrator’s skill set should determine the level of support needed.
Desktop environment
refers to the graphical interface where the user does her work. Complete desktop environments consist of a large number of tightly integrated but separate pieces of software. By default, GNOME uses a window manager called Metacity, and KDE uses kwin. The window manager is, of course, the most important part of the desktop. It handles window placement, movement, appearance, and user interaction with all the windows on the desktop. The choice of window manager is probably the most important factor in the success of a Linux migration on desktops.
Several window managers are available that offer a wide spectrum of options in the tradeoff between resource use and functionality. Table B-1 illustrates the point. XFCE4 is a less resource-intensive desktop environment. It is also highly customizable.
Table B-1. Window managers and resource usage
Desktop |
Required RAM |
Required CPU |
---|---|---|
XFCE4 |
128 MB |
100 MHz |
GNOME 2.x |
384 MB |
800 MHz |
KDE 3.x |
512 MB |
1 GHz |
In examining and evaluating desktop environments, consider the following enterprise-user requirements as paramount:
Ease and consistency of use. An intuitive, integrated, and consistent experience for the end user is key.
Ease of migration from Microsoft Windows.
In examining productivity software, we confine ourselves to word processing, spreadsheet software, and presentation software, since meaningful alternatives to Microsoft Office exist in these three categories. We consider the following as requirements for enterprise users:
Ease and consistency of use (i.e., an easy and integrated experience; e.g., the ability to create compound document spreadsheets or images embedded in a document)
Ease of migration from Microsoft Office
Interoperability with Microsoft Office document formats
Note that open source productivity software such as Open Office is steadily gaining ground on major proprietary platforms such as Microsoft Windows and Mac OS X.
Applications in this category are alternatives to Microsoft Access, Microsoft FoxPro, and Adobe FileMaker. While local applications using a desktop database are fast becoming a rarity in enterprises, we recognize the continuing need for an alternative to Microsoft Access. Most open source productivity suites provide alternatives to Access by providing a usable interface to existing open source database servers such as MySQL or PostgreSQL.
Applications in this category are alternatives to Microsoft Explorer. The security problems and lack of regular functionality updates in Microsoft Explorer have created an opportunity in this area, and the Mozilla Project has capitalized on it. Mozilla Firefox is a modern, stable, and secure web browser available for all major desktop environments.
Since messaging is a collaboration tool, we provide a detailed explanation of messaging clients in Appendix D.
3.133.108.68