Why Support GNU/Linux?
"If only sitting was required, all frogs would be
Buddhas" |
I probably do not need to outline the virtues of the famous open source applications. You all know Linux, Apache, Sendmail, Perl, and you all know how these have changed the face of business as we know it. This much you know, but what you may not know is how the very essence of these projects, the principles behind Open Source, are changing the ground rules for business itself. It began with a famous offspring of the open source movement, the Internet.
ClueTrain Manifesto:Networked markets are beginning to self-organize faster than the companies that have traditionally served them. Thanks to the web, markets are becoming better informed, smarter, and more demanding of qualities missing from most business organizations.
To traditional corporations, networked conversations may appear confused, may sound confusing. But we are organizing faster than they are. We have better tools, more new ideas, no rules to slow us down. [CTM]
A new culture prevails in the business world. We live in an economy where the primary resource is not limited and regulated, but, freely exchanged, it multiplies as it is traded. "Mergers and acquisitions" do not warn of multinational dominance in a globalized economy. These changes are no more than an Industrial mindset grappling with the problems of global communications; not realizing the Grand Unified Corporation already exists (it is called "Humanity"), they expand their castle walls in a fool's quest to streamline communications. I wish them luck. There are no "internal" communications.
Communications is Everything Open source is not a revolution. It is an evolution, a recognition that knowledge in motion fuels our economy, that knowledge is only additive and its value accrues only when it is shared. The FSF could demonstrate the potential of a free exchange of ideas, but it took a critical mass of open communications to complete the picture: Plain talk freely moving between peers is the secret progenitor behind the famous offspring of the Internet like Apache and the Linux O/S. If "Open Source" did not exist, it would not have been necessary to invent it, it would have happened. |
When the prospective Open Source Consultant asks "Why should I support GNU/Linux?", this could could be read, "Why should I add support for GNU/Linux products to my repertoire?". I propose a deeper meaning. I propose defining support in the political sense, in the cultural sense of supporting the social forces which are changing the computing landscape, and changing the basis of commerce.
How Big is Open Source?
Let's take a short aside to re-assure those who want to know if it is worth their while to service the Open Source marketplace. Linux may give some clues as to the size and rate of growth of this movement. Different pundits put projections on the growth of Linux at different rates, but almost all agree: We are headed for world domination, perhaps within the next 3 to 4 years. Linux already commands 31% of all webservers, 24% of all servers, and 5% of all desktops, and this only considers those units which were sold. (see Figure 1)
Figure 1. IDG Press Releases
IDG and other industry pundits almost universally find Linux as the only O/S growing in market share, and growing at an alarming rate
The Growth of Linux
"How do you extrapolate 'Exponential'?"Jeff Prothero's 1998 essay called "The Last Dinosaur" asked "If Linux has been doubling its market share every six months for the past decade, how should an analyst extrapolate the curve?". When his paper was first posted, Linux was estimated to command 2.5% of the market; Jeff projected 10% of all desktops by 2000 if the doubling trend continued, followed by 40% by 2001 and market saturation (aka world domination) by January 2002.
Our most conservative estimates put us a little behind those figures, but his essay accurately predicted how the unthinkable could happen in 1999 when big business discovered a business model (or many business models) for Open Source. Linux became inevitable: Once just a few the large enterprises came on board, others were keen to follow (see Figure 2. Jeff's essay concludes ...
"The conservative assumption, then, is that when Linux reaches 10% of desktop market share, the third-party desktop software developers will likewise jump on the bandwagon. Understandably: a 10% market share gain is enough to interest the shareholders. We may expect to see this avalanche effect to kick in on the desktop market sometime in the year 2000."
Figure 2. IDG: Charting the Use of Linux
Microsoft's new Windows 2000 is as much a bonus to Linux after its delivery as it was during the long delays.
More telling than sales and other market metrics of size and growth are the Internet "mind share" metrics for Linux. If we can agree that business will be deeply intertwined with the Internet, then a measure of the buzz over open source may be useful. Table Table 1 is a simple and unscientific experiment comparing relative search hits on AltaVista. Over just 14 months we see every O/S losing percentile market share except Linux, with some slight gains in other Unix systems once the Linux Fever began in early 2000. What is abundantly clear is how, among all its peers, Linux is rapidly gaining the mind share of the Internet. Remember what I said about good ideas on the FidoNet?
Table 1. Comparative "hits" for AltaVista based on [JP]
| Keyword | Jan 10, 1999 | Oct 29, 1999 | Mar 10, 2000 |
|---|---|---|---|
| Windows | 2,530,775 | 6,748,317 | 9,997,800 |
| Linux | 502,053 | 984,844 | 5,876,340 |
| Solaris | 251,513 | 453,884 | 586,690 |
| HP/UX | 105,833 | 180,612 | 214,587 |
| FreeBSD | 81,781 | 180,547 | 671,895 |
| MacOS | 70,851 | 133,882 | 161,925 |
| UnixWare | 23,386 | 32,994 | 85,415 |
| Ultrix | 15,133 | 26,351 | 38,530 |
| OpenBSD | 11,892 | 13,354 | 52,960 |
Industrial Time Lags
Figuring aside, mass acceptance of new technology doesn't just happen out of statistical necessity. Are we closing in on the days of Linux in 'prime time'? I believe it is inevitable, but before you rush to buy more shares, I'll add a caveat: For very real reasons which may even be natural laws, world domination may not happen as soon as we'd like. Figures like Netcraft's Web Survey tell us that, in certain markets, Linux and Open Source are already the industry norm, and they hold great promise for the future, but there is still one more principle to consider: You cannot rush evolution.
Bucky Fuller, famous inventor of the geodesic dome, spent a lot of time charting the time lag between invention and mass acceptance. He discovered fixed lag times specific to each industry. In electronics, it is 18 months. In the building trades, 30 years.[RBF]
Rather than be discouraged, Bucky concluded an inventor should build their better mousetrap even if the initial response is dead silence. If it is truly useful, mass adoption will come, but there is no way to hasten it; it comes in its own time (I hope Michael Cowpland is listening!). It is important only to realize that process can not start until the inventor announces the new invention.
Time-lags in Software Engineering
My own informal observation in our industry shows we are no where near as fast as the electronics industry. We took 4-5 years to accept CPM, 4-5 years for DOS (just in time for the Mac), 4-5 years for GNU-C and Emacs, 5 years for Windows 3.1, and it has taken just under 5 years for Windows95 (it was introduced in 94 and was still pretty fringe in 1998).
Linux was introduced as a research tool in 1991. By 1996, Bob Young found Linux commanded a fair mind share of that market. Bob introduced the packaging of Linux for the server/developer market in 1996, and he's right on schedule with his IPO. Mandrake scored product of the Year in 1998 which puts the capture of the desktop market at 2003, and Corel's bid to capture the office desktop close behind in 2004. It's kinda spooky how these dates fit both the IDC pundits and Jeff's growth-doubling algorithm.
There are two lessons here. The first is an illustration how the "Release early, release often" credo only gives the illusion of hastening mass adoption by pushing ahead the initial announcement, and the second is that, for those of us vying to service this industry, these dates give us our time lines. If we want to know what is feasibly to have widespread by 2003, we need only look back two years in the Linux support industry for our clues.
Milestones to World Domination
At the ITAC "Roadwork" conference in 1994, I was talking with an old-timer about the soon-to-be ubiquitous Internet. At the time, I thought we would see mass acceptance of the Internet very quickly, 1996 at the latest, yet I was continually frustrated by the blank stares I'd get when I offered to put a business online, or even install free email for them. My colleague gave me an observation which proved itself then, and which proves itself in every new technology. He said:
When I was a young boy, no one in their right mind would own an automobile. To own a car, you had to be part carriage-maker, part small-motor repairman, part bicycle repairman, and be a little suicidal. Before we could have a car in every driveway, we had to have a mechanic in every village, gas stations on every corner; if you car was broken, either someone could fix it quickly, or they could lend you another.
Open Source consulting is about that infrastructure. It is about putting a mechanic in every village, about roadways and driving lessons and all the infrastructure an industry needs to survive. Eazel could put the next killer desktop out on the market tomorrow, or Cobalt could release the killer network appliance, but without the corner shops to sell and service them, without the books, technical support and guidance in every village, they won't go anywhere.