Where McNealy's wrong about open source
- 02 September, 2002 10:32
I found interesting comments in an interview Robert McMillan conducted with Scott McNealy. Let me sum up McNealy's views, and what I think is right and wrong with them. While it's not my intention to alter a subtlety in McNealy's argument, please send me a message if you believe I have, and include an explanation of where I went wrong.
Point 1: Open source screws up revenue modelsMcNealy believes open source is screwing up Sun Microsystems Inc.'s revenue models. The result is that Sun does not have enough money to market Sun ONE to the same degree that Microsoft Corp. markets .NET. The result could be a total win for .NET.
I can't deny that open source is screwing up revenue models, and not only Sun's but also the revenue models for many companies. It is also difficult to refute his contention that this cripple's Sun's ability to market Sun ONE effectively. I would like to say there is a sense of balance because open source is interfering with both Sun and Microsoft's revenue models. But Sun doesn't have the billions in the bank that Microsoft has accumulated over the years through an abuse of monopoly power, so Sun is at an unfair disadvantage.
At a personal level, I admit some sympathy for McNealy's views. In the first place, I'd hate to see Sun ONE suffer at the hand of either Microsoft's marketing bucks or the economic effects of open source, because I happen to admire the Sun ONE vision and the technologies behind it. My sense of fairness says Sun should get some benefit from the R&D efforts it has put into J2EE and all of the other technologies it developed. Worse, if .NET wins because open source undermined Sun's revenue model, that would not only be bad for Sun, but for the open source community, as well.
More important, however, is that I would like to see someone -- Sun or anyone else -- engage in an effective marketing campaign that exposes the misinformation behind the .NET campaign. It just so happens that Sun is the company most motivated to do this, since it has a reasonably complete alternative to .NET, Sun ONE. Sun ONE is, as McNealy states elsewhere in the interview, based primarily on open standards, open protocols and open interfaces, so it is far less likely to lock you into Sun as a vendor than .NET locks you into Microsoft. As alternatives to both .NET and open source go, Sun ONE has a lot going for it.
The problem I have with McNealy's complaint is that he's making it with blinders on. It's entirely possible that McNealy isn't telling the whole story here, or that if he is, it will be revealed in the complete interview. (The article includes only a portion of this interview.) Scott assumes Sun has only one way to make money, and open source is gumming up those works. The flaw in his reasoning is that there isn't only one way to make money.
Yes, open source is screwing up conventional software revenue models. However, nobody is holding a gun to McNealy's head to force Sun to stick to conventional revenue models.
IBM Corp. was probably the first company to reinvent itself around a viable model for the future. Ironically, IBM took its first step in this direction when it gave up on some of its software products (such as OS/2 and SmartSuite), endorsed Java, and started promoting open standards as the only reasonable course of action. As a former user and fan of OS/2, I resented IBM's move back then, and I resented even more that it was induced in part by Microsoft's refusal to let IBM license Windows 95 for a reasonable price unless it put the brakes on OS/2.
In retrospect, however, this could have been one of the best moments in open source history. IBM's transition to Java and open standards eventually led to its support of Linux and open source, steps it probably wouldn't have taken if IBM had held onto the dream of supplanting Windows with OS/2 and Office with SmartSuite. We can also credit IBM for given Linux a great deal of credibility by endorsing it. In a twisted way, we can thank Microsoft's hard-core monopolistic practices for much of the success of Linux today.
The point is simple. Where was Sun when IBM was making all these changes? Sun should be well into its plans to redefine the company to compensate for the fact that these models not only must change, but have already changed. If Sun is now struggling to find a new revenue model because it held onto the hope that it could beat Windows with Solaris for too long, whose fault is that but Sun's?
I don't mean to criticize everything Sun did. After all, IBM had to endorse Java -- IBM didn't invent it. However, the critical mistakes Sun made are easy to identify, at least with 20-20 hindsight. Sun never should have licensed Java to Microsoft, and should have invested whatever was necessary to create its own Java virtual machine for Windows.
As hard (impossible?) as it would have been to sell this idea internally, Sun should have come out as one of the first to adopt and endorse Linux, and market it at least as aggressively as it marketed Solaris. The best technology isn't the one that always wins, and Scott McNealy admits that in this very interview. When internal dissent at Sun undoubtedly revolved around the superiority of Solaris, the answer should have been that Linux was going to succeed whether Sun liked it or not, and Sun would be in a better position in the long run if it was perceived as one of the first to recognize this trend.
Is it too late for Sun to endorse open source and Linux? Is it too late to begin the transition to a more services-based revenue model? I don't think so. There are rumors that Sun is about to go through a major restructure, so perhaps that is an indication that Sun is perfectly aware of the changes taking place and is knee-deep in its plans to adapt. We'll see.
Point 2: Who cares about an OS?
My favorite point Scott makes is no one cares which OS runs on the microchips in your car, cell phone, or most other appliances. He argues it's only in the narrow world of IT where people care about operating systems. Sun ONE is an attempt to make the OS distinction go away, because Sun ONE runs on any platform. If you develop to Sun ONE instead of an OS, you don't need to care what OS is running underneath.
That's a terrific sales pitch for Sun ONE, and I agree with the philosophy and approach. Sure, it's self-serving for McNealy to promote Sun ONE as a better alternative than developing for Windows or Linux. Sure, this is as much about money as it is ideals. Sun is at least as interested in undermining Microsoft and Windows as it is in providing the customer with the ideal development platform.
You can't deny this approach does, indeed, benefit both developers and customers, at least for the near future. There might come a day when a victorious Sun might try to wield monopoly power as ruthlessly as Microsoft has. I doubt this could happen, but it's certainly possible.
I endorse the concept, even if Sun isn't involved. As much as I am committed to the Linux platform, I would love to see the day when people stop thinking in terms of Windows, Linux, Unix, Solaris, or any other operating system. No single company deserves the blame for the sad state we're in today. IBM, HP, Sun, and others didn't really market their OS separately, but they did splinter Unix to promote proprietary hardware.
There's no way to measure the damage Microsoft did by marketing DOS and Windows as if it were a software product equivalent to a word processor or database. In any other industry, this concept would be insane. It's one thing to buy accessories for your car or buy a new car, but can you imagine what it would be like if Ford tried to sell you a new engine every year?
Getting off Scott free
I find myself sympathetic to Sun and McNealy. Sun creates great technologies, and Sun endorsed open standards and platforms before "open" was hip. Sun even released most of the source code for Star Office, and inherited open source Mozilla.
Nevertheless, Sun missed opportunity after opportunity to trumpet its achievements, and it has missed several opportunities to establish itself as a hero in the Linux world equivalent to IBM or better. There's still time, but that time will eventually run out. I'll be watching.