The Technofile Web site has moved.


Technofile is now located at http://twcny.rr.com/technofile/
Please update your links, bookmarks and Favorites.  
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
 

Microsoft's dilemma: Fix Windows or give up trying?
technofile  by al fasoldt

Columns and commentaries in a life-long dance with technology
Simple gray rule


Microsoft's dilemma: Fix Windows or give up trying?


Technofile for Sept. 12, 1999

By Al Fasoldt

Copyright © 1999, The Syracuse Newspapers

The changes sweeping the PC industry this year are hard to follow. They can make you dizzy.

They all seem to center around Linux. But don't be fooled. What's happening in the world of PCs is as much about Microsoft as it is about the upstart operating system with the funny name.

Linux is hot for three basic reasons: It's new, it's ultra-reliable and it's free. Microsoft, which makes Windows, is in trouble for three equally basic problems with Windows: It's unreliable, it's expensive and it's broken.

Let's take a look at each of these reasons in turn.

First, the reasons Linux is hot.

It's new.

There hasn't been a new operating system for personal computers in a decade or more. Millions of twentysomething computer users have known no other operating system than Windows. To people who get their kicks from computers and software, the arrival of an advanced new operating system is a breathtaking time.

It's ultra-reliable.

Linux isn't just reliable. Toasters are reliable. Grandparents are reliable. Linux is in a different category. It's sun-baked-and-brick-hard reliable. It's designed to keep running no matter what happens. It's engineered to sail on smoothly even if programs start to misbehave. If Program A crashes on a Linux computer, Program B is totally unaffected. And the operating system itself remains cool and calm. Windows users sometimes have a hard time believing all this; they're used to PCs that crash and programs that cause other programs to fail. They're used to the Microsoft propaganda that has tried to convince all of us that crashes are just a part of everyday PC life. Linux users know this slightly differently: They're a part of everyday Windows life.

It's free.

You can buy Linux if you want to get the operating system packaged up in a pretty box, or you can get it free off the Internet. If you don't like big downloads, you can order it by mail without a pretty box for a couple of dollars. This is the hardest part of the Linux saga to explain. How can something so good be free? And how come you can pay for it if in fact it's supposed to be available without charge? Linux is good because it's constantly fixed and improved. (This happens daily.) Something that good can be free because the people who created Linux want it to be free. They believe good software should be free. (They believe bad software should be free, too, in fact. No doubt they'd tell you that bad software just withers away in such a setting because nobody would choose to run it when good software is available for the same cost nothing.) And you can pay for it even when it's free because many companies package Linux their own way, charging for the customization or just the printed manuals.

In the cost of computer maintenance, Linux is hardly free, of course -- somebody has to fix hard drives that fail or modems that stop working and that kind of thing, regardless of the kind of operating system -- but it's about as inexpensive as computer support can get. Because Linux doesn't crash or lock up, and because Linux PCs don't get themselves messed up so badly that they can't even get themselves running, support for Linux PCs is not marred by the high labor costs of tracking down Windows conflicts and trying to fix them.

Look at it this way: Non-working PCs are not an option for any business or organization, so somebody has to get PCs that crash back up and working. Because Microsoft can't solve the problems that Windows PCs have -- Windows is the cause of these problems, so of course Microsoft can't fix them -- the technical support people have to figure out fixes themselves. And those fixes usually do not work for long, because Windows is not reliable enough.

And now the reasons Microsoft is in trouble.

Windows is unreliable.

People who are in charge of computers at large companies should be laughing at this. Or crying. Windows 95 and Windows 98, the two main modern consumer versions of Windows in widespread use in late 1999, cannot be trusted to do the five main tasks of an operating system. They can't run for long periods without crashing, can't control programs that misbehave, can't protect the operating system from unintentional (or intentional) meddling by users, can't deal properly with the PC's memory and can't keep their own system files in a stable configuration. There are many other flaws, of course, but those are the five that tell you Windows can't be trusted.

Windows costs a lot of money.

Windows costs about $100 if you buy it in a store. (The "upgrade" version might cost less and the full version might cost more; the average probably is $100.) Manufacturers pay less than you or I do when they buy Windows to put on new PCs, and, of course, they pass the cost on to their customers. Windows in such cases might cost $50 to $70. But that is not the real cost of Windows any more than the $2.75 you spent for a razor is the real cost of shaving. The real cost of Windows is staggering. It shows up in the cost of Microsoft's supporting software (its server programs, office suites and so on), in the cost of support personnel to maintain Windows PCs and in the almost incalculable cost of lost wages and productivity caused by problems that plague Windows. A computer that can't do the job it's supposed to do is worse than a useless computer; it's an expensive useless computer.

Windows is broken and can't or won't be fixed.

When your 11-year-old drops Grandma's china serving dish on the floor and it breaks into a zillion pieces, you know exactly what he means when he mumbles "I'll fix it." He means he's sorry. And of course you know he'll never fix it. It's not fixable. That's the state Windows is in. It's not fixable. Partly this has come about because the code in Windows is just too badly written to be fixed, and partly it's the result of Microsoft's almost complete misunderstanding of what matters to all of us. Microsoft has never shown that it realizes how bad Windows is and has never taken on the responsibility of actually fixing it, and this makes many of us distrust Microsoft's motives. If you need an example of how Microsoft does things without showing any real interest in doing what its customers need, consider this: If a serious bug is found in Windows, Microsoft posts a fix for Windows (a patch for the code) on its Web site while continuing to sell the bad version in stores and to PC manufacturers for months and months. In case you blinked and missed the meaning, let's restate it as simply as possible: Microsoft continues to sell versions of Windows that it KNOWS are bad. If GM did this with cars, we'd have the Big Two automakers instead of the Big Three.

Imagine what this means. Microsoft makes an inept operating system for PCs. There's no real competition, so PC buyers have no choice. We get Windows, like it or not. If all mid-size cars came with the same lousy engine, the situation would be the same. We could choose the brand of car and the color and the interior trim, but we'd all be stuck with the same bad engine. If there were no other engine manufacturers, we might even learn to accept our fate and try to put up with the inadequate engine, forgiving it for all its faults. Likewise, we've lived with Windows and learned how to sidestep some of its oddities. We've spent a lot of money on software that tries to fix the worst problems of Windows. We've tried hard to believe the company that makes Windows; we've tried hard to believe that Windows is the best thing that ever happened to computers, that Windows will make our lives better, that Windows will become the operating system for microwave ovens and TVs and VCRs and even watches, just as Microsoft tells us it will.

We've tried hard. We all did, except for a few crazies who insisted all along that Apple's Macintosh computers were better. Of course, as we now realize, those crazy people were right, but they were right for the wrong reason. They were right because Windows is the swiss cheese of operating systems, not because the Mac was some sort succulent feast. Windows has been smoke and mirrors. Even a Mac that is turned off does a better job at some things than a Windows PC does. (It's more stable, for one thing.) It's no contest. All through the useful life of Windows, every possible competing PC operating system has been better in one way or another. OS/2 from IBM was much more reliable, GEM from Digital Research handled memory much better, GeoWorks from the company with the same name had a far better interface. Yet the Mac never got past 12 to 15 percent of the market, and all the possible rival operating systems failed to threaten Windows. Microsoft knew how to advertise, knew how to get what it wanted and knew how to put its own products into computers in almost every home and office. Microsoft was an earthmover, and we were Windozed.

It's standard human psychology to accept what you cannot change. We could not survive otherwise. In the useful life of Windows, we can identify three phases. Windows was growing and improving in the first phase, was stagnating in the second and is declining irreversibly in the current phase. By the middle phase, we had begun to learn the lesson that Microsoft was unwittingly teaching us that Windows fails, that Windows costs too much, that Windows is poorly supported but we had no choice of operating systems. There was nothing else we could do. We accepted what we could not change.

Linux pushed Microsoft into the third phase. Suddenly we were able to see the emperor clearly, and we could see that he had no clothes. Nothing that Microsoft says seems believable any more. An operating system that cannot perform properly is not acceptable any longer. We have a choice. For years, we accepted what we could not change. But change is easy now. The finery the emperor told us he was wearing is missing. The emperor has no clothes.

And so we find ourselves in the middle of a revolution. We find ourselves choosing the way our computers will behave, the way they will run. It is an easy change, and, for many, a quick change. Learning a few new ways of opening and closing programs and spending an hour studying new commands is easier than spending all weekend every six months of so trying to get a computer running again. Making choices is easier than living with regrets.

This is the lesson that Bill Gates, the $100 billion founder of Microsoft, will learn over the next 12 to 24 months. he and his company have to make some choices. It's a tough time. Linux will not go away and Microsoft cannot buy it, since it is not for sale. (It never will be for sale, because it is an Open Source operating system, available to everyone without charge.) Since it can't be bought, Linux won't go away. To Bill Gates and his increasingly shrinking band of followers, Linux is like the unwanted guest. It won't leave.

So the first choice Microsoft has to make isn't really a choice. To someone standing on a burning deck, the choice of whether to jump into the water isn't actually a choice. You jump or die. To survive, you don't need to worry about making a choice. You just jump.

The next choice Microsoft has to make is whether to scrap Windows entirely or try to fix it. Scrapping Windows would be an admission that Windows is a flop, but it's a much smarter course than trying to fix Windows. Fixing Windows is impossible, since the code is too inscrutable to fix. Once a programming project gets beyond hundreds of thousands of lines, software engineers are no longer able to understand how all parts of it work separately and together. Windows has many millions of lines of code.

Microsoft could create a new operating system patterned after Linux or it could even create a new version of Linux itself -- everyone is free to do this -- but that option would force Microsoft to make its own code public. The computer source code used to made changes to Linux must be made available to anyone who would like to change the changes, if you follow me; Open Source software is constantly being improved and bugs are continually being fixed because programmers around the world jump in to help fix problems and improve functions that aren't working right. Microsoft would have to meet that standard if it came up with its own version of Linux, and that means Microsoft would have to publish the source code for its new operating system. And that would mean, of course, that Microsoft's new operating system would have no secrets, no hidden code that gives Microsoft programs an advantage. It also means a 14-year-old in Topeka could improve it and post the improved version on the Internet, or a university professor could add a function and make it available on the university's server computer, or a rival company could repackage Microsoft's new operating system with a few improvements.

It's a tough choice. Microsoft probably won't reengineer Windows. It knows as well as we do that Windows is not fixable. It probably will create a completely different operating system that will run Windows software without exhibiting the problems Windows has. It might call the new operating system "Windows" (as in "Windows 2002," maybe) but that would be a mistake. Windows has a bad reputation and that means it has a bad name. Microsoft probably is already busy negotiating to buy Trumpet Software, an Australian company that has come up with an operating system that runs Windows programs without the flaws that Windows has. Whether Trumpet is for sale or not isn't known, but the operating system it has developed could be worth hundreds of millions of dollars now and many billions of dollars a few years from now.


 Image courtesy of Adobe Systems Inc.technofile: [Articles] [Home page] [Comments: afasoldt@dreamscape.com]