accesses since September 18, 1998
copyright notice
link to published version: Communications of the ACM, October, 1998
accesses since September 18, 1998Remember the Mosaic War? It was the hot topic of techie conversation just a few years back. The term hearkens back to the kinder and simpler era of Web antiquity (circa 1994!). Like navigator/browser, helper app X-windows, the term signifies a bygone era - the Web gilded age where every software developer thought they had a chance at market dominance and Web surfing was a favorite pass-time. It might be useful at this point to see if we can identify winners and losers in this Mosaic War of old, especially if we could then anticipate the outcome of remaining hostilities. But first we wander down memory lane.
The Web was conceived by Tim Berners-Lee and his colleagues at CERN (now called the European Laboratory for Particle Physics) in 1989 as a shared information space which would support collaborative work. Berners-Lee defined HTTP and HTML at that time. As a proof of concept prototype, he developed the first Web client navigator-browser in 1990 for the NeXTStep platform. Nicola Pellow developed the first cross-platform Web browser in 1991 while Berners-Lee and Bernd Pollerman developed the first server application - a phone book database.
By 1992, the interest in the Web was sufficient to produce four additional browsers - Erwise, Midas and Viola for X Windows, and Cello for Windows. The following year, Marc Andreessen of the National Center for Supercomputer Application (NCSA) wrote Mosaic for X Windows which soon became the browser standard against which all others would be compared. Andreessen went on to co-found Netscape Communications in 1994 whose current browser, Netscape Navigator, succeeded Mosaic as the next de facto standard Web browser. That's when the Mosaic War started - aggressive and imaginative developers trying to out-Mosaic Mosaic. The feverish pace of development was something to behold. Within months, literally dozens of new start-up companies appeared. By 1994 it appeared as if browser vendors would proliferate like wire coat hangers.
FIGURE 1: Browser by name by year
Fast forward to 1995 (see sidebar) - a turning point in the Mosaic War. By year's end, Mosaic is basically out of the picture (see Figure 1) as a navigator/browser. It went from over 90% of the browser market to under 5% in just over two years. So the first major fatality in the Mosaic War was, ironically enough, Mosaic itself. Meanwhile, Netscape displaced Mosaic as the de facto standard within the same time frame and became the new de facto browser standard. By the end of 1995, spirited developers worldwide were all attempting to capture the half of the browser market which wasn't already Netscape's. By this time the Mosaic War became known as the browser war for obvious reasons.
So, in the end, Mosaic went the same way as Cello, Viola, Erwise, and Midas before it. These Web fatalities confirmed that the Web was highly unforgiving of technology deficiencies. On the other hand, the big winner of the Mosaic War was clearly Netscape. Netscape's dominance was the result of a constant stream of innovations -- much to the chagrin of the World Wide Web Consortium (www.w3c.org) and the Internet Engineering Task Force (www.ietf.org) which preferred to introduce innovations in an orderly and deliberative manner through their RFC's and standards committees. In any event, some of Netscape's more popular innovations appear in the table, below.
Armed with imposing innovations in 1995-7, Netscape appeared to be the clear victor in the browser war as well. However, two external factors changed Netscape's future. First, the dominance of Windows as the OS of choice for the overwhelming majority of Web users (see Figure 2) provided a strong disincentive to developers of client-side software for other OS environments. As Windows rapidly became the dominant OS, Netscapes commitment to multi-platform development (approximately 20 platforms) became increasingly uneconomical. While the expenses of Web client development are basically constant across platforms, the potential revenue streams vary with the size of the customer bases - i.e., the Windows market is approximately 20 times the size of the MacIntosh and Unix markets, and hence potentially 20 times the revenue. The commitment to multi-platform development really hurt Netscape's overall competitiveness, since their main rival, Microsoft's Internet Explorer chose to focus their development effort primarily on a single-platform.
FIGURE 2: Primary platform - market share by year
The second major event which challenged Netscape's hegemony in the browser arena was Microsoft's combined marketing strategy to both 1) provide Internet Explorer without charge, and (2) bundle it with Windows OS. Not surprisingly, Netscape found that it is difficult to compete with products which are both seamlessly integrated into the operating system and also free. This situation was not overlooked by the Department of Justice who recently brought an antitrust case to the US Court of Appeals claiming Microsoft has used its operating system dominance to achieve a monopoly in the browser market. Netscape has since unbundled its browser, Navigator 5.x, from its groupware suite, Communicator, and has agreed to release and license the source code of Navigator to interested developers in a desparate attempt to keep their code alive.
Figures 1 and 2, taken together, show that Internet Explorer is doing to Netscape what Netscape did to Mosaic. There is an important difference, however. Netscape unseated Mosaic primarily through innovation, whereas Microsoft's successes are primarily due to its dominance of the OS market and its un-rivaled marketing prowess (I assume that few would claim marquis, background sound, table colors and stationary backgrounds are strong representatives of IE innovation.). Failing intervention by the Federal Trade Commission, or successful antitrust litigation against Microsoft from the Department of Justice, it appears that history will repeat itself in the most recent incarnation of the browser war. It seems all but inevitable.
So what's next? The end of the browser war coincides with the beginning of the desktop war. The stage is set: Goliath Microsoft will do battle with the Davids of client-side software development over compatibility with the full range of Windows applications. The war is heating up as we write as the major players try to outflank each other on choices between Windows API vs. Swing Set interfaces, Visual Basic vs. Java scripting languages, dynamic HTML vs. pure HTML with Java document standards, Secure Sockets Layer vs. Java security model, ActiveX vs. CORBA middleware, and so forth. This fight is going to get nasty before its over (sometime around the turn of the millenium, I suspect). We're talking trench warfare here with digital nerve gas.
But the previous wars will pale in comparison to the ultimate bloodbath, over embedded applications in which the forces of good and evil will fight for supremacy over our thin Web clients: PDA's, televisions, appliances, automobiles, phones, games, smart cards and digital jewelry (that's right, Sun has already created a prototype of a Java ring! - cf. www.javasoft.com). The embedded apps war will be the true test of mastery over things digital, because it will extend dominion beyond the general-purpose computing desktops to the special-purpose embedded applications around which our life is based. This topic is so new that developers are just now beginning to define their positions. By the time that the winners of the desktop war are identified, the embedded applications war will be in full swing.
So, let's return to our original question: "Who Won the Mosaic War?". The answer is Netscape. However, as we've seen this is a shallow victory for Netscape as it struggles to hang onto its leadership position in the client-side browser area.
As the dust settles on the browser war, the apparent victor is Microsoft's Internet Explorer. However, like the Mosaic War, the outcome of the browser war may not have much strategic impact, because of its narrow scope.
In terms of impact, the effect of the desktop war will be far more important. While Microsoft remains in the leadership position, the absence of widely accepted standards makes it possible for real innovation to surface. Smart money will probably bet on Microsoft to emerge victorious. But even the desktop war won't by itself have the most effect on the world of networking. That honor will go to the big war, the no holds barred, winner-take-all embedded applications war which will be fought over our televisions, air conditioners and security systems. The embedded applications war will be the Bosnia of the Internet software development community - and could easily lead to digital ethnic cleansing as developers scramble among chaos for survivability. Unlike previous wars, the embedded applications war will be about lifestyle computing - control over the digital appliances which we take for granted.
I'll repeat what I said in 1995. These are exciting times!
Online preprints of his article quoted above may be found at http://berghel.net/col-edit/cybernautica/jan-feb96/pcai961.html. Preprints of related articles and columns may be found at berghel.net/publications/ and berghel.net/col-edit/.
Well, among other things I was writing articles and columns on the Mosaic War. Remember that at that time Windows 95 had just been released, and Microsoft's Internet Explorer remained in beta. Here are some of my observations at that time about a possible Microsoft Web Monopoly? - which is perspicacious given recent Department of Justice actions against Microsoft. (for online references, see berghel.net/col-edit/cybernautica/jan-feb96/pcai961.html and berghel.net/publications/mosaic-wars/os2-mag.html
(1) According to the most recent (at that time) World Wide Web user surveys, the client side of the Web changed from an almost exclusively Unix to environment to a primarily Windows environment in just eighteen months. The percentage of survey respondents who used Unix went from 88% in late 1993, to 44% in late 1994, to 9% by April, 1995. At the same time, the percentage of users who used Mosaic as a client navigator/browser went from 97% to 58% to 3%.
(2) At that time, the distribution of host operating systems was 9% Unix, 26% Macintosh and 52% Unix.
(3) Also at that time, the distribution of Web clients went roughly as follows (numbers are % of total)
| IBM WebExplorer | 1 |
| MacWeb | 1 |
| AIR_Mosaic(16bit) | 2 |
| Lynx | 2 |
| NCSA Mosaic | 3 |
| NetCruiser | 4 |
| PRODIGY-WB | 27 |
| Netscape | 54 |
The question I asked is whether these numbers were worrisome.
There were some obvious trends identified in this data. For one, the operating system prominence of Windows shows that the use of the web was now in the hands of the hoi polloi. The high priests of Webdom who conceived of, and implemented, the Web represented and ever-shrinking fraction of its overall use.
Second, the domination of Windows in the enormous SOHO market gave it unparalleled advantage over other browser and OS vendors. I predicted that by mid-1996 it would be extremely difficult to compete in the Web browser arena because of the overwhelming control exercised by two products: Netscape and Microsoft's Internet Explorer, despite bravura performance from a variety of browser developers.
I predicted by 1997 that it might be hard for even Netscape to retain market dominance. I suggested that it remains to be seen whether even Netscape Corporation could leverage superior technology into a strong enough marketing position to withstand any attempt by Microsoft's to control the Web client markets as it had done with PC operating systems. I saw in the recent Justice Department prosecutions of Microsoft a certain deja vu.
As I said in '95, this will be an interesting year for the Web.