Post of the Day
May 20, 1998
Subject: Re: Break up MSFT to increase value
Well, MicroOS might still be a virtual monopoly, but MicroApps would probably have a rougher time than you envision -- once they lost their ability to exploit undocumented system calls.
I envision the opposite: In the exceedinly unlikely event of a simple division-by-division breakup, MSApps would still have a virtual monopoly and the MS/OS group might have a rougher time -- once they lost their ability to exploit a fully-developed application test-bed.
|"Think about it. Excel and Word, the flagship Windows applications, were actually first written for the Mac OS, and ported to Windows much later. Microsoft apps ruled on the Mac before Windows existed."|
So much for an advantage from unfair hooks into OS code. In fact, this "common wisdom" is exactly backwards. Often as not it's been MSFT's apps division that writes "undocumented" calls that eventually make it into the operating system.
For instance the Excel team cooked up DDE, the precursor to OLE and ActiveX, because Windows lacked a method for inter-app communication. Similarly the WinExcel and Winword development teams wrote the hypertext-like context-sensitive Help applet because the online help developed by the Windows team caused problems for users and developers alike. These and countless other innovations came from elsewhere in the company and were later published and incorporated into Windows.
We've speculated in previous threads why Windows has prevailed over it's competition. I propose that MSFT with it's Apps, Languages, and Networking divisions was able to use it's in-house applications developers as real-world test centers for their operating system development. And that the reason IBM, Sun, and others have not prevailed is that they lack similar insiders driving their OS extension and improvements. Apple, when it had it's Claris division, had a similar advantage that might explain why it's still number two.
I'm going to hypothesize why: If applications aren't developed for an operating system, and if hardware vendors don't install it, that OS isn't really going to go anywhere. Therefore the OS has to answer more to hardware vendors, application developers, and MIS departments. And these guys tend to be boat anchors on innovation and streamlining. (Example: why does every keyboard still have a SysReq and Break key? Why wasn't MSFT able to one of those as their Windows key?)
|"Smart developers don't rely on one OS for that reason. They certainly won't depend on "undocumented" hooks into one OS for their entire advantage, either."|
On the other hand, consumers buy whatever machine and OS runs the applications they need. App developers don't have to satisfy hardware vendors, OS developers, or language developers. They have to satisfy end users (yes, and MIS for corporate accounts but even then the emphasis will be more on productivity than hardware compatibilty.) If the OS gets in the way of the application, developers and customers will scream. And if necessary the developers will jump to the OS that best serves their need to satisfy the customer.
One last case in point: Corel was able to develop successful apps for Windows. They certainly would have no trouble doing it for the Mac or even OS/2. That they failed to do so in JAVA suggests� what? The problem there isn't just too many undocumented hooks, it's too few documented ones. MSFT, with it's experience developing applications, is able to see this. Sun, with it's lack of experience, isn't. MSFT's solution? Add deeper support for apps. Sun's solution? Sue MSFT.
Smart developers don't rely on one OS for that reason. They certainly won't depend on "undocumented" hooks into one OS for their entire advantage, either. MSFT applications developers are smart. In the event of a breakup I'd worry less about MS-Apps than about MS-OS. (Note: I still wouldn't worry much.)
What do you think?