October 2000

Break Up Microsoft, Are You Kidding Me?

BY Tom Keating

More important, however, are my convictions about the importance of application portability, and the value of standards -- even de facto standards -- in fostering third-party development. Application portability is bound to be important to communications solutions, for not only do PCs act as communications appliances, they will also prove indispensable as administrative windows (pun intended), at least for end users who would benefit from a user-friendly means of configuring their own communications services.

PC FROM THE START
Almost eight years ago, I graduated from the University of Connecticut with a B.S. in computer science and engineering. Right out of college, I assumed programming and IT support duties for TMC. I set up the corporate LAN and e-mail system, and I upgrading the PCs, many of which were still using monochrome monitors.

I soon learned which department in TMC was the most demanding -- the art department. They wanted Macs on their desktops. But Rich Tehrani and I said, "No, we're a PC shop!" We didn't relent, even though the art department needed desktop publishing tools, which were far superior on the Mac than on the PC in 1993. We felt that the PC was the best choice for the company. Most of our business applications were designed for the PC, and we weren't eager to accommodate two operating systems (OSs). That would have brought on more headaches and training and support issues.

I TOUCHED A MAC ONCE, JUST ONCE
I wasn't anti-Mac in 1993. In fact, I had extensive experience using a Macintosh, since my college roommate, who was also a computer engineer, had one. We used the Mac for word processing and designing newsletters and greeting cards. My roommate used to scoff at my NorthGate 286-Mhz PC, calling it "un-user-friendly."

My NorthGate lacked a graphical OS (such as Windows 3.1) because I didn't have enough room on my 60-MB hard drive. Sure, my PC had only 1,024 KB, and it started with a black DOS screen, but I liked it, though I would note wryly that I was king of the command line. (Eventually, I had the last laugh. On several occasions, our courses forced my roommate to use my PC. There was the time we took an assembly language course, and there were a few electrical engineering courses that used PC-based programs such as PSpice.)

RELIEF IN SIGHT
When Microsoft released Windows 95, I knew that the Mac's days were numbered. Did I have any regrets? Well, I would have liked the release to have come a little sooner. For two years as TMC's chief MIS support technician, I had put up with all the headaches of supporting both the Mac and the PC. While TMC didn't have any Mac PCs, several customers did use Macs. Also, the service bureaus that printed our magazines all used Macs. One problem I encountered was dealing with fonts when going cross-platform from the PC to the Mac. The kerning differed, resulting in document reflow and more than a few cuss words from the art department as they shouted, "We want Macs!"

Another problem was that every time someone sent an article to TMC on a Mac-formatted disk, we couldn't read it. I had to use purchase and use a conversion utility to read the disk, and if memory serves me correctly, it could read only low-density Mac floppies, not the high-density Mac disks. I often asked people to either use a low-density disk or to simply send the document in PC format.

But that was not the end of my travails. Even when I was able to read the Mac disk, they would send the file in Macintosh Word or another word processing format that our PCs might fail to read! And so I cried out to the OS gods, "Please let there be just one OS on the desktop! These inter-OS support issues are killing me!"

MIRACLES DO HAPPEN
Just when I said that, I heard a whisper in my ear that said (to quote from Highlander), "There can be only one ..." My prayers were answered when Windows 95 was released. Virtually overnight, Windows 95 became the number one desktop OS, which caused my support issues to drop dramatically.

SAY IT AIN'T SO
Some people would like the DOJ to break up Microsoft, creating several "Baby Bills." They argue that Microsoft has a monopoly and therefore impedes competition and innovation. And they cite the example of the Modified Final Judgment, saying, "Well, look what happened when the government broke up AT&T. It helped produce more telecom companies, more competition, and lower rates for the consumer."

In my opinion, these arguments are flawed. Let's compare the analogy to AT&T. Okay, I agree that breaking up AT&T did more to spur the telecommunications industry's growth than any other recent government action. However, there is a big difference between the breakup of AT&T and a breakup of Microsoft (should that occur).

Some anti-Microsoft sentiment is due to a misconception that an OS monopoly is inherently bad. Well, consider what might happen if we were to equate OSs to phone services. Granted, in the post-modification world, I have my choice of phone services from WorldCom, Sprint, SNET (SBC), AT&T, and other phone carriers. But I know that from carrier to carrier, the analog, T1, and ISDN offerings will be more or less consistent. I can be assured that I will be able to pick up my phone and get dial tone once I terminate the trunk lines from any of the carriers into our corporate PBX.

Now, let's look at OSs. Suppose I have my choice of OSs, and I decide to install Linux on my PC. (For purposes of argument, let's say that installing Linux is the equivalent of getting an ISDN, T1, or analog line to my office). Of course, just having an OS isn't enough. You need applications to run on the OS. Perhaps I want to buy Quicken and run it on Linux. Sorry, Quicken wasn't programmed for Linux. Okay, let me try Need for Speed III, a popular game. Sorry again, not designed for Linux. Are we seeing a trend?

I should point out that you can get at least one popular PC game on Linux -- Quake. However, most companies lack the resources to program their applications for multiple OSs, so they focus their development efforts on Windows. To examine this analogy from a graphical perspective, please see Table 1.

Table 1. Technology Comparison

The key thing I want to point out in Table 1 is that it is technologically impossible for every application to run on every OS. Yes, many applications can be ported from one OS to another, but that involves engineering resources and time. If there existed 20 popular desktop OSs, that would require 20 different versions of the software to be coded. So a single, dominant OS isn't necessarily "bad," and holding a grudge against Microsoft simply because it has a dominant OS is foolish.

THE DE FACTO STANDARD
The Microsoft OSs (Win 95, 98, NT, 2000) are similar enough that third-party vendors may port applications from one OS to another without too much difficulty. Fortunately, the four OSs share enough core components that third-parties are spared undue programming effort.

For better or for worse, Windows owns the desktop, becoming the de facto standard, which is convenient for third-party developers. Imagine if Intuit had to program a version of Quicken for Windows, Sun, Linux, HP-Unix, SCO, Macintosh, and 20 other OSs. Intuit would have to quadruple their engineering staff to be able to program their financial software to run on several disparate OSs. In addition to the Windows version of Quicken, I believe Intuit has programmed their software for the Macintosh since the Mac still has respectable desktop market share. Thus, it was worth Intuit's engineering effort to spend some R&D dollars, since they would recoup their investment.

But what if many competing OSs were to arise? Let's suppose that in the future, Microsoft loses about 65 percent of its desktop market share (bringing it down to 20 percent) and that four other competing OSs also each own an equal 20 percent of the desktop market. Now, Intuit would have to hire five times the number of developers -- hiring programmers who are specialists in each OS. Of course, this isn't exactly true, but regardless, as a programmer myself, if I write a program that runs on Windows, I know for sure that it isn't going to run on Linux. I have to rewrite the code virtually from scratch, especially the graphical window API calls, which differ immensely from one OS to another.

Getting back to my analogy, T1, ISDN, and analog are "standards" which allow competition to flourish in the telecom world. I hate to say it, but the only equivalent standard in the computing world is the Windows OS, which last I heard held over 85 percent of the desktop market. Breaking up Microsoft to help other OSs gain desktop market share is foolish. That's like saying "Well, there are too many T1 lines out there and we don't like it. I think we should force the telecom carriers to come up with a new T1 standard." Does anyone remember the ISDN nightmares where several competing "flavors" of ISDN caused headaches galore?

THE REAL ISSUE
Of course, the real issue isn't whether Microsoft has an OS monopoly (that's hardly disputable). And it isn't even whether an OS monopoly is desirable. The real issue is whether Microsoft unfairly leverages its OS advantage in other realms.

In my view, Microsoft's alleged abuses should be balanced against the competition Microsoft has enabled in other realms. There is plenty of room for competition on the server side as well as for developers, who may write third-party desktop applications on the Windows platform. Sun Microsystems, for example, makes excellent servers for the Central Office/telecom space, as well as very good Web servers. Fortunately, there are more standards on the server side. For example, I can run SMTP (e-mail standard), a Web server (HTTP standard), a DNS server, or a firewall using Microsoft, Linux, Sun, or several other OSs. They'll all get the job done.

And yet another factor encourages competition on the server side -- you don't need a "standardized" GUI. As long as the IT guys have a stable, robust platform, they don't care about nice, pretty GUIs. This isn't true on the desktop side. Having a "standard" GUI (Windows GUI conventions) helps collaboration and communication between fellow employees. It also saves IT personnel time in training. If you know Microsoft Excel, you can probably navigate Microsoft Word without asking IT personnel for much help.

PROGRAMMING UTOPIA
To further illustrate my support for non-Microsoft OSs on the server side, I should mention that we use Linux server within TMC for several functions. For example, it acts as a firewall and as a DNS server. Linux is a lot cheaper, requires much less hardware horsepower than Windows, and in some ways is easier to implement than Microsoft's equivalent.

I suppose if all OSs were 100 percent Java compliant and all applications were written using Java, then applications could, theoretically, run on any OS. But even the "open source" Linux OS isn't designed using Java. Most of it is written using C and C++. If we lived in a "Star Trek world," we could just send any software application through the "universal translator," and the computer would magically convert the application to run on any OS. Sorry to burst any Trekkie's bubble, but Star Trek is fiction, not fact.

Sure, Microsoft cheated to become the desktop king. But that doesn't mean we should cause a fragmentation of the OS market. Just look at the fragmented UNIX community (not to mention my nightmares supporting the Mac), and you can see why having standards of some sort is essential. I'm all for competition, but I sincerely hope the DOJ doesn't create a situation in which several competing OSs end up vying for the desktop market. I don't think this is the DOJ's intention. Rather, it is my belief that the DOJ is trying to stop Microsoft from abusing its power, indulging in questionable business practices, and strangling competition unfairly.

ALL IS NOT FORGIVEN
I completely agree with the DOJ that Microsoft should be punished, but I disagree with the punishment (namely, breaking up Microsoft). I should point out that I am a big fan of Netscape, and I think what Microsoft did to Netscape was reprehensible. I eventually switched from Netscape to Internet Explorer (IE) because I liked IE's History and Favorites features, and because IE crashed less often than Netscape. The fact that Microsoft first gave out IE for free and then embedded the browser into the OS was anti-competitive. Moreover, Microsoft included a shortcut to IE on every Windows desktop. Considering this, I agree with the government that Microsoft virtually destroyed Netscape's market share via questionable tactics. Another less-than-noble Microsoft episode: Using strong-arm tactics, Microsoft prevented many computer manufacturers such as Dell from including any competitive browser on the desktop.

HYPOTHETICALLY SPEAKING
What if Microsoft were broken up? What would happen? Instead of getting a Dr. Watson error once in a while on my Windows 2000 computer, I'd get them eight times a day! Here's why: Right now, if the Microsoft Windows Media development group has buggy code that crashes Windows 2000, they can simply consult with their colleagues in the Windows 2000 development group, and say, "Hey, we found a bug. Here's a copy of the stack dump." Various Microsoft departments can communicate and collaborate to fix the bug.

If the DOJ gets its way, these departments would reside in different "Baby Bills," forbidden to communicate with each other. Any collusion between these separate "Baby Bills" would no doubt result in some sort of legal action. I don't understand how preventing Microsoft from communicating with various essential programming teams is in the best interest of the consumer. If anything, this will stifle innovation and make for "buggier" Windows code that will eventually hurt the customer.

I have to agree with Microsoft that breaking up the company will hurt innovation for third-party developers -- but especially for Microsoft. No one can argue that Windows hasn't improved workforce productivity, especially since inter-office and even inter-corporate communications is much easier with a standardized OS on the desktop.

A BETTER SOLUTION?
In the future, Microsoft should be prevented from dictating exclusive agreements that squelch competition. I propose that an oversight mechanism be put in place, so that contracts between Microsoft and outside parties are subject to review. A confidential committee, appointed by the courts, could detect any anti-competitive provisions.

SENTIMENT ASIDE
I think what gets lost in all the anti-Microsoft sentiment is the plain fact that Microsoft has provided a means by which third-party software developers can inexpensively develop an application that can reach a mass audience. I should also point out that software written by third-party developers vastly outnumbers any applications written by Microsoft. Microsoft can have the measly $89 that I paid for Windows 98. Guess how much money I have spent on third-party applications that run on Windows? We're talking thousands of dollars spent on applications such as Photoshop, Tomb Raider, Unreal, Quark, Quake, Quicken, and others. So just because I run Windows 98 on my home PC doesn't mean Microsoft is making a dime off of me, other than the $89 I paid for the OS.

As long as Microsoft doesn't muscle third-party developers out of the market, then let Microsoft have the desktop space. An example of where I would draw the line would be if Microsoft offered Microsoft Money for free with every copy of Windows in an attempt to take market share from Intuit's Quicken.

I keep using Quicken as an example. Would you like to know why? Because Quicken makes superior financial software to its closest competitor, namely Microsoft Money. Hmmm. So the best product was the one I chose. Isn't that what a democratic, free market society is all about? How American. DOJ: Keep your eye on Microsoft, punish them if you must, but I hope that common sense will win out. I hope you will realize that breaking up Microsoft is not in anyone's best interest. 

[ Return To The October 2000 Table Of Contents ]