Behind the masks of Microsoft and Sun

Okay. Fair enough. There is in this quote, however — as there is in much of the carefully scripted testimony delivered by Gates, Scott McNealy (Sun Microsystems), and Jim Barksdale (Netscape) — a distinction that can only be described as Clintonian. “There is no improper relationship,” said the president of his current imbroglio (with unquestionable veracity due to the use of the present tense). “I don’t want to be labeled as whining,” said Jim Barksdale before the Senate, carefully describing what he wanted as opposed to what he appeared to be doing.

The software industry is starting to look like the ancient Greek theater, in which actors spoke their lines from behind masks, each mask designed to illustrate a theme. However, behind these masks lie real issues, and some of them are difficult to comprehend. The definition of monopolistic practices, like that of sexual harassment, is, to some degree, in the eye of the beholder. The edge cases are easy to see and understand, but the middleground is often blurred. There is always history to rely on for clarity, however.

Swept up in the tornado

History is always visible in the software industry. Old dogs may learn new tricks, but they still perform the old ones best. If you doubt that statement, read Geoff Moore’s excellent book Inside the Tornado (Harper Collins, 1995), the follow-up to his popular book on high-technology marketing (Crossing the Chasm). Compare his description of Microsoft’s competitive strategy in operating systems to its current actions in the Internet space. For you Microsoft bashers out there, Sun’s strategy provides an equally relevant and occasionally humorous case study.

Moore’s premise is that a dominant player will emerge in any high-tech market segment over time (Microsoft in operating systems, Oracle in databases, Intel in microprocessors, and so on). That company will represent at least 50 percent of the revenue and an even greater percentage of profits in the market. Companies that win in this environment exhibit one of several competitive strategies. Moore outlines Microsoft’s (and Intel’s) winning strategy on the desktop as follows:

1. Recruit partners to create a powerful whole product
2. Institutionalize this whole product as the market leader
3. Make the product a commodity by designing out your partners

Moore points to Intel (microprocessor), Seagate (disk drive), Novell (network software), WordPerfect (word processing), and Lotus (spreadsheet) as the partners that allowed Microsoft DOS to dominate its market and institutionalize the Wintel platform as the market leader. Once DOS was perceived as the default operating system, Microsoft used Windows to eliminate its software partners one by one.

WordPerfect and Lotus were late to market with Windows versions of their software and were quickly replaced by Word and Excel. Novell is still a major player in networking software, but clearly under pressure from Microsoft (especially with Windows NT). As a result, most personal computers today are configured with a Microsoft-only software environment consisting of Windows, Word, Excel, and PowerPoint (and often Access). Only Intuit’s Quicken is close to being another standard piece of this commodity environment. (And Microsoft would have happily acquired it, had it not been for the intervention of the U.S. government.)

These actions are all legal, of course, and in many ways admirable. They have created a company that a critic like Sen. Orrin Hatch (R, Utah) refers to as a “gem” and Jim Barksdale calls a “national treasure.”

We find the same strategy being executed on the Internet with even less reliance on traditional software partnerships. Web browsers and HTML authoring are part of the new environment — give the first away for free and acquire the second and price it aggressively. The result? Internet Explorer is blowing past Netscape Navigator/Communicator as the default browser on Windows, and Front Page has a dominant marketshare. People want access to entertainment and information? Sign exclusive relationships with major information providers like the Wall Street Journal and Disney and limit their ability to promote competitive software from Netscape.

What about Java?

And then there is Java. Here the traditional partnering model is being played out in full. Partner. Institutionalize. Eliminate your partners. History repeats itself. Or does it?

The curtain has yet to fall on the first act of the Java melodrama, but two things are certain, and three questions remain. First, the market wants a standard, cross-platform Java. In the absence of this Visual J++ for Windows is being bought up and is rapidly becoming an institution. The three major questions are obvious:

1. How will the legal action between Sun and Microsoft resolve itself?
2. Will Java ever achieve its “write once, run anywhere” potential?
3. What do I do in the meantime?

The answer to the first question may be coming soon. Federal Court Judge Ronald Whyte heard preliminary arguments in the Sun/Microsoft case on February 27 and took under advisement a Sun motion for a preliminary injunction that would stop Microsoft from using any Java-compatible logo with Internet Explorer 4.0. Microsoft’s counsel David McDonald claimed that his company had the right to use the logo because its products had passed earlier test suites. Later tests, he said, were not backward-compatible. Sun is claiming that Microsoft knew that Internet Explorer 4.0 failed the compatibility tests but used the logo anyway. Sun claims continued use of the logo is causing confusion in the market and doing irreparable harm to the Java platform.

The second question will take longer to answer than the first. The road to “write once, run anywhere” will be a long one. How about “write once, run anywhere with only a few, very minor modifications”? It may not be as catchy as the first slogan, but it would be a major improvement over today’s options.

The third question has several unsatisfying answers. Developers can write to the Java 1.0.1 specification, which both Microsoft and other vendors support. They can carefully try to sort through the incompatibilities of the competing SDKs and APIs and try to develop cross-platform code that is somewhat more sophisticated. Finally, they can try to pick a winner and write to that specification. These are all difficult choices. Most developers seem to be treading the middle path. (See the November edition of JavaWorld for some advice on how to do this — in John Zukowski’s “How to avoid pitfalls of non-standard Java”).

Conclusion

In the end, what does all this mean? Simply that Microsoft is executing what has always been a successful strategy. The difference this time is that its very success leaves it open to charges of monopolistic practices and increasing scrutiny by the U.S. government and the governments of other countries.

Is Microsoft a monopoly? That is for the courts to decide. Jim Barksdale clearly has an opinion. In a theatrical flourish before the Senate hearings, he turned to the audience and asked anyone who used a personal computer to raise his or her hand. The room was filled with waving arms. He then asked anyone who was using a PC without Microsoft’s operating system to raise a hand. All the hands fell. “Gentlemen,” he said to the Senate panel, “that is a monopoly.”

Most observers are predicting that the U.S. Department of Justice (DOJ) is preparing a wide-ranging case against Microsoft’s business practices — from its tying arrangements between Windows and Internet Explorer to the exclusive agreements it has written with the group of content providers for its push channels. In response, Microsoft has made its contracts with 40 ISPs and computer vendors less restrictive. It is also reported to be considering killing the channels on Internet Explorer. So far, Java seems to be absent from the list of technologies being considered for litigation by the DOJ.

With Visual J++ 6.0, Microsoft takes its version of Java to the threshold of becoming a commodity product (in the Geoff Moore sense). Visual J++ 6.0 provides excellent support for Microsoft BackOffice, integrates with the latest incarnation of its component model DCOM+, and provides an enhanced set of class libraries that give access to native resources (Windows Foundation Classes). The Wall Street Journal reports (March 12) that the DOJ does not have plans to block Microsoft from offering a Windows 98 version that includes Internet Explorer. Therefore, it appears likely that the Microsoft virtual machine is also on the threshold of becoming a true commodity.

How the Java suit will conclude is anyone’s guess. As predicted here in my previous column, it is likely that Sun will win at least a part of its suit. Interpreting all the ramifications of the licensing agreement between Sun and Microsoft undoubtedly will keep the court occupied for some time. However, another prediction from last month is also likely to be true: It is doubtful that the resolution of any current litigation will force Microsoft to ship a version of Java that it does not wish to ship.

In the meantime, what is the poor Java developer to do? Sit back and enjoy the theatrics. There are still many acts to come. It is sometimes hard to know which mask to cheer for and at which mask to hiss, but this is clearly the best show in town.