September 16, 1996 When it comes to judging Microsoft products, the devil is in the details In the Bible, in John 8:44, it says of the devil: "When he lies, he speaks his native language." Yes, this column is about Microsoft Corp. But don't worry. I'm not one of those who think Bill Gates is the devil. I simply suspect that if Microsoft ever met up with the devil, it wouldn't need an interpreter. My most recent example surfaced when Microsoft delivered final Windows NT 4.0 code to InfoWorld. In that meeting, I probed Microsoft about the differences between Windows NT Server and Workstation to investigate the Web server issues raised by Netscape Communications Corp., O'Reilly & Associates Inc., and others. I asked specific questions and got specific answers: There is no way to change any setting to make the Workstation kernel behave like the Server. The reason, said the Microsoft representative, is that the source code for the kernel has embedded statements -- #ifdef statements. These cause the compiler to produce different executables depending on whether the target is a server or a workstation. As a result, the two kernels are hard-coded to use different caching algorithms and multitasking priorities, among other things. That's what the fellow said, in front of a room packed with InfoWorld editors and analysts. And it simply isn't true. Microsoft will undoubtedly argue that the difference between what it claimed in the meeting and what is true is simply a matter of semantics. It's not. Ask any programmer. It's also extraordinarily easy to disprove. Windows NT's own file-compare program will show that Workstation and Server kernels are identical, bit for bit. O'Reilly & Associates' Andrew Schulman points out that the reason the two kernels behave differently is due to two settings in the Windows NT registry (see http://software.ora.com/news/ms_internet_frame.html). The system makes attempts to prevent you from changing these settings, but it's pretty easy to get past the guards. If you do, you'll break the law, and Windows NT Workstation will perform just like Windows NT Server. Self-destructive behavior I find this terribly depressing, but not because it may have implications about what you are paying for and what you get. It bothers me because Microsoft is still engaging in damaging propaganda. And, although I'm not sure Microsoft understands this, the target is itself. Lies invite harsh criticism, reinforce Microsoft's reputation (deserved or not) as an evil empire, and give Microsoft's enemies ammunition. They also undermine every press release and public statement. Each time Microsoft addresses a controversial product feature, a slew of analysts and hackers will try to disprove its claims because they know that, more often than not, they'll score. Worst of all, Microsoft's intent for giving misinformation about controversial issues often goes well beyond the limits of reason. To paraphrase an ex-Microsoft employee, Microsoft would have been better off proclaiming from the rooftops that there is DOS in Windows 95 rather than trying to sustain the deception that there isn't. The predisposition to lie also draws suspicion in every questionable situation. For example, a few weeks ago, a Digital Equipment Corp. representative demonstrated limited Plug and Play and power-management services for Digital notebooks. According to this fellow, Microsoft, along with Digital and two other vendors, created a set of undocumented hooks to the NT kernel to give OEMs the opportunity to fill these gaps in the OS. I wondered how many other vendors had access to these hooks, so I posed exactly that question to Microsoft and a few other OEMs. The OEMs I spoke with were either unaware of these undocumented services, or they said the issue was too sensitive to talk about on the record. After stonewalling for nearly two weeks, Microsoft denied that any such hooks exist. (Microsoft declined to speak to InfoWorld news reporter Carolyn A. April as well. See "NT laptop vendors to add Plug and Play," Sept. 2, page 6.) Did Microsoft innocently provide hooks to only a few vendors and afterward realize how stupid it was to limit the selection? Were these intentional favors to a few OEMs? Are there really any hooks at all? I honestly don't know. But I probably wouldn't even be asking these questions had it not been for an established pattern of deception. After all, how do you give Microsoft the benefit of the doubt when you know that if you throw it into a room with truth, you'd risk a matter/anti-matter explosion? Worst of all, the tendency to deceive erodes the credibility of almost everything you read about Microsoft. How many of you disregard Microsoft articles for this very reason? Of course, Microsoft is not the only guilty party. But when you consider that most of the industry runs on Microsoft products, the potential damage these deceptions can cause threatens to be far-reaching. For now, I'll leave it to others to ponder how damaging this particular lie is to Microsoft, as well as whether or not Microsoft's differentiation between its NT Server and Workstation constitutes anti-competitive behavior. What bothers me more is that it's a silly and unnecessary lie -- like one spoken from a company for whom lies are a native language. You can send me e-mail at nicholas_petreley@infoworld.com. Honest. To respond to this column go to the forum. Missed a column? Go back for more. Copyright © 1996 by InfoWorld Publishing Company