To Ian Murdoch and the Open Solaris Community, As a long-time follower of the open source community, I stumbled across this forum a week ago and was intrigued by the activity taking place on Open Solaris. I spent some time reading posts, and over the day some concerns became clear to me; concerns that I felt compelled to write about. I have written a letter that I hope will encourage discussion and bring about positive results. I thank you for your interest.
[b]Gaining a significant market share will attract developers. [/b] Imagine with me that Solaris was the OS on 40% of all computers. Do you believe that it would have a large and vibrant developer community? Most certainly. Developers would miss out on 40% of the market if they ignored Solaris. Getting Solaris installed on as many computers as possible should be the first and main concern. It is true that users alone will not make Open Solaris into a great product, but they are the final judge of its success. The Open Solaris community must remember who will use their software when the development is done: the user. If a large number of desktop users should be the first priority, targeting the Linux developer’s community is a poor strategy. These developers are already a small subset of all developers, and hundreds of Linux distros compete for their limited time. In targeting them, Solaris has chosen the hardest possible target when much easier targets exist. Where, then, will this imaginary 40% of users come from? From Windows, of course. Windows is where it must come from. Solaris [i]must [/i]take a large share from Windows, not Linux, to be successful. Fortunately… [b]There is a large group of Windows users who would use a FOSS alternative if it existed.[/b] Wait. What about Linux? We already have an open source alternative. Not really. There are two kinds of people who currently use Linux. There are those who are good at programming, who understand computers and software at a deep level. These people enjoy the command prompt, know what “Vi” is and usually work in network administrator or other IT positions. I shall refer to these people as the Technical Elite. The Technical Elite use Linux. The second group is the friends and relatives of this Technical Elite. Their moms and dads and roommates. These people rely upon the Technical Elite to help them get Beryl working, to figure out Grub, to re-compile the kernel to make their wireless card work. They are not power users, and use their computers to simply surf the web and write email. The only reason this second group can even consider Linux is because it has become much easier to use. A lot of work has been done to hide the power, so that an average Joe can do most average things. But should something go wrong, should something advanced need to be done, anyone not a Technical Elite must seek the aid of one. There is no middle ground. The arcane syntax of the command prompt ensures this. In fact, I predict that Linux will continue to grow until the Technical Elite can “support” no more. Linux is too hard for the average user to fix or update on their own. There are too many gotchas, too many places where special knowledge is required. If any one doubts this, a simple visit to the Ubuntu support forum will confirm it. Post after post documents the trials and tribulations of average users wrestling with even the (seemingly) simplest of tasks. Most people use Linux for a single reason. They use it because, at some level, they agree with the principles of Open Source. They take the time and energy required to learn Linux because of the strength of that belief. They are like people who have a dream of climbing Mt. Everest. They train and practice and spend a lot of money to fly to Asia and risk their lives to realize this dream. It is an expensive and dangerous process, but the view from the top makes it all worthwhile. However, for every person who has the time and money to spend making their dream come true, there are hundreds of others who share the same dream, but lack the time and money. Climbing Mt. Everest will always remain a dream for millions. It is the same with Linux and Solaris. It is a wonderful dream that thousands like myself have, but we lack the time and money to invest in the learning required to make it happen. I want the freedom that climbing Mt. Solaris offers, but I have a job and a wife and debts and not enough free time to spend years figuring out sudo or XFS or D-Trace. A truly Free operating system will always remain just a dream for me. I, and thousands like me, use Windows because we want more power than a neutered Linux, but we lack the knowledge of the Technical Elite to make Linux work. It is unfortunate, because it doesn’t need to be this way. The… [b] Technical constraints that shaped the way we use computers no longer exist and must be rethought.[/b] Some here remember a time when programs were punched in to cards. We have come a long way since then. There are still some relics of that era that continue in computing, particularly in open source communities. Developers have spent many years acquiring these skills. They have grown used to the command prompt, to Perl scripts. They are comfortable. Their knowledge gives them status. However, these relics are the chief hurdle that prohibits the mass acceptance of FOSS platforms. Until the command prompt (and all functions that rely on it) is replaced by a graphical, intuitive, easy-to-use interface, open source operating systems will never gain wide acceptance. The barrier of entry is simply too high. The frustrating part is that it doesn’t need to be this way. We now have the technology to make the command line obsolete. Combo boxes and check marks and tool tips could open up a world of features to the general user. Simple, intuitive interfaces would also allow the best tool for the job to be chosen, instead of the most familiar. Making it simple enough for the non-Technical Elite makes it simple enough for all. Outdated and arcane software ideas that have been kept on life support for 30 years could be retired in favor of ideas that are the product of 30 years of learning. An operating system, based on excellent technology, wrapped up in intuitive interfaces that allow the average user to utilize the full power of their computer would be extremely attractive to not only the current Linux community, but the much larger community of users who are attracted to the ideas of Open Source. For the thousands of us who will never climb Mt. Everest, it would be a dream come true. [b]Open Solaris has a once in a lifetime opportunity.[/b] Solaris is uniquely situated to make this dream a reality. With the support of Sun Microsystems, resources (no matter how insignificant) could be directed towards making fundamental aspects of Solaris “powerful AND easy” for all users: to replacing the command line with an interface that would allow many more users to enjoy and maintain their own computers, to rethinking the way people interact with their computers, to lowering the barrier to entry without resorting to baby-sitting. Should Solaris use sudo or RBAC? Users don’t care about which is better. They want the one they can use (Windows’ dominance demonstrates this), and neither of these technologies currently meet this criteria. An operating system founded on Open Source ideals that allowed users to easily master their computers would be extremely attractive to a large group of people now laboring in Microsoft’s camp. The much desired developers would come of their own accord, should Solaris gain a significant majority of users. Worry about the technology being used, but more importantly, worry about making the technology easily usable. In your debates and meetings and focus groups ask the question, “Could a complete newbie use this in less than 5 minutes?” Keep forefront in your mind that every time the answer is “No”, the barrier to entry just got raised a little higher, Solaris’ market share just got a little smaller, and a few more developers just picked a rival operating system to develop for. Picture a future where Free Solaris is the world’s most popular operating system. Looks good, doesn’t it? The only way this can be is by appealing to users. Remove the barriers that keep this from becoming reality. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org