A new Linux group has been organized to look at user interfaces and help formulate a standard in an effort to ease the way for Linux to move onto the desktop.
by Randy Jay Yarger
Over the past year, public exposure of Linux has increased beyond anyone's expectations. Company after company has embraced Linux as a superior solution for server applications. Despite all this, one mantra of negativity has continued to dog Linux: ``It's not suitable for desktop use.''
Even when extolling the virtues of Linux as a server, the praise is often followed with a few discouraging warnings. Linux is difficult to install. The command line is too cryptic. The X Window System's applications have no consistency. Linux doesn't work with the major Windows and Macintosh applications. Linux is only for engineers who like long strings of numbers. (I've actually heard this last one, believe it or not.) They all boil down to ``Linux is too hard for the average user.''
While many people, even Linux users, do believe the above, for each assertion against its ease of use, there is a group of people who know it's not true and can prove it. Anyone who has installed Windows from scratch can testify that it is hardly a simpler process than installing Red Hat or Caldera OpenLinux. Likewise, anyone who has attempted to move a set of files from one machine to another using standard DOS commands can appreciate the sheer usefulness of tar, gzip and other commands.
In these cases, as well as many others, the simple fact is that other systems seem simpler only because most people don't encounter these situations when using them. When you buy a new PC, Windows is already there, pre-installed. Most common file operations are performed using point-and-click and drag-and-drop, not with command-line utilities. For years now, most versions of UNIX have offered similar ways of manipulating files in a quick and easy graphical manner.
Where do these misconceptions about Linux originate? Like most things in life, people believe what they see and what they hear. We know what people are hearing about Linux, so what do they see? Typically, a person's first encounter with Linux is from a friend or colleague who is an established Linux user.
Most users who have spent a considerable time with Linux have learned the ``Great Secret of UNIX''. The one feature that UNIX (and Linux in particular) has which no other operating system has been able to duplicate is its infinite configurability. Every moment of the Linux experience, from startup to shutdown and everything in between, can be tweaked, tuned and completely subverted to create a uniquely personal environment.
People who see Linux for the first time invariably encounter a system that is efficient, streamlined, and to their eyes, completely incomprehensible. The same features that make Linux attractive to those who know it, keep those who don't, away.
So what is the solution? How do we take the one system that really can be all things to all people and present it in an intelligible way to newcomers?
The answer may lie, at least in part, in a new movement known as the Linux/UNIX Independent Group for Usability Information (LUIGUI). Dr. Nathaniel Borenstein of the University of Michigan announced LUIGUI at the inaugural meeting of the U of M branch of SIGCHI (the Computer and Human Interaction group of ACM). Dr. Borenstein's name is familiar to many in the computer world. Among many other accomplishments, he is the inventor of MIME, the recognized standard for identifying and transferring data over the Internet. Dr. Borenstein has also been an active participant in the push for responsible Internet development, an author and an entrepreneur.
LUIGUI is an attempt to formulate the ideal interface for those new to Linux. Taking the project to its ideal conclusion, anyone trying Linux for the first time would be guaranteed several things.
Intuitive installation: one of these days, you will be able to go to your nearest Best Buy or Sears and purchase the latest Compaq or Sony PC pre-installed with the industry standard desktop operating system: Linux. Okay, I hope that will happen. However, until it does, a PC will come with Windows, and anyone who wants to try out Linux will have to install it. Therefore, the process should be simple and painless, even enjoyable. This is an essential step for Linux. It doesn't matter how many impressive demonstrations their friends have given--a person's opinion of an operating system is solidified the first time they sit down and use it on their own.
Full functionality: Linux, at its very heart, is simply the name of an operating system kernel. Linux provides the backbone of a useful system, but no utility on its own. What we commonly call ``Linux'' is actually the Linux kernel in combination with a wide variety of tools and applications. In fact, the sheer number of standard Linux utilities created by the GNU project alone has led to suggestions that the full Linux distribution be named ``GNU/Linux'' or even ``LiGNUx''. While that can of worms is best left unopened, what is undeniable is that the quality of applications available for Linux has been as much a factor in its popularity as the quality of the kernel. Taking this idea further, there is a core set of functionality that every user needs, or at least should be able to access. Word processing, a spreadsheet, image manipulation--these are all tasks that have become everyday activities in the world of desktop computing.
Ease of use: this area is the key for many people and is where much of the current development work (outside of LUIGUI) is being focused. Projects like KDE and GNOME are proving that, yes, Linux can have a GUI that the average person can use comfortably. This is the ``pretty face'' Linux desperately needs. Users should be presented with an initial desktop that is simple, elegant, and above all, understandable. It should take minimal thought to launch an application, navigate the file system or do any of the other common tasks that users perform through the desktop.
Ease of configuration: this is an area where Linux presents two extremes. On one hand, the configurability of Linux is unmatched by any other mainstream operating system. This is very important, as having a user-friendly configuration system is not much good if there is little the user can configure. However, in the presentation of configuration, Linux is still lacking. Users should have a method of customizing their environment that is as easy to use and understand as the environment. The fact that Linux has so much that can be configured makes this task even more difficult, but considering the feats the Linux community has been capable of so far, it should not be beyond reach.
Standardization: this is, for LUIGUI, both the hardest and the most important area. All the above criteria can be met, in some form or another, by Linux today. New, easy-to-use installers are becoming common on the major distributions. Company after company has announced applications for Linux users, and the Open Source community is matching their output. The desktop environment is friendlier then ever, and graphical configuration utilities are appearing for everything from adding users to configuring high-end web servers. However, what every new Linux user must be guaranteed is an environment that is predictable. A single, standard desktop environment, meeting all of the criteria above should be presented when a Linux system is booted for the first time. Naturally, people will still customize their environment, but if this standard is as good as we know it can be, a large enough base of users will form so that this desktop environment will be used as a sort of lingua franca with which all Linux users can communicate.
At first glance, accomplishing the above seems like a near-impossible task that would take millions of dollars of funding and years of programming talent. However, Dr. Borenstein approaches this task with a surprising mantra: ``Zero lines of code.'' While not meant literally, since new tools will inevitably have to be written, his statement points to the heart of the matter: almost everything we need is already present.
``People have been trying to improve the user interface for UNIX for over 20 years, and every single attempt has succeeded. It's like shooting fish in a barrel'', says Dr. Borenstein. The tools and configurability UNIX provide make this task technically simple. However, no consensus has ever been reached on what the best standard interface would be for new users and all users in general. The one graphical system that has become standard, the X11 windowing system, explicitly made no choice on what the standard interface should be and merely provided the tools to create new interfaces. This is where LUIGUI comes in.
Dr. Borenstein has called LUIGUI the ``Consumer's Report'' of Linux. His intent is that the members of LUIGUI will evaluate all of the various user-interface options currently available for Linux. Through methods used by human-computer interaction professionals and computer interface design specialists, it should be possible to take the various pieces that exist now and combine them to form the ``ideal'' standard interface for Linux. After individual pieces of the interface are evaluated, the best of the crop will then be advocated by LUIGUI to the rest of the Linux community. If enough public trust and support is placed in the project, then hopefully the major Linux distributors will make some effort to implement the standards proposed by LUIGUI.
The LUIGUI project is spearheaded by graduate students in the University of Michigan School of Information. The self-described ``project cheerleader'' of LUIGUI is T. Charles Yun, a master's student in the U of M's SI program. Yun, along with web-site manager Katherine Degelau, created the LUIGUI home page at http://www.luigui.org/. A list of the current LUIGUI projects and goals can be found there.
The LUIGUI project is still in its formative stage, and help from all areas of the Linux community would be appreciated. Instructions for joining the LUIGUI mailing list can be found on the Web at http://www.luigui.org/community/involve.html. Members of the project feel that with broad support throughout the Linux community, we can once and for all put to rest the last of the Great Myths about Linux, and finally enable it to make inroads among desktop computing users.
Randy Yarger (randy@yarger.tcimet.net) is the Systems Administrator and Chief Programmer for H-Net, Humanities and Social Sciences On-line (http://www.h-net.msu.edu). His sign is Virgo, his favorite color is blue, he wants to work on world peace and is currently practicing for the interview portion of the Miss America pageant.