Microsoft: Doing Everything But

This essay was originally authored in September of 2000.

Microsoft (or in the lexicon of many dissidents: Micro$oft) has been a computing power for years. They, under the leadership of their founder Bill Gates, may probably be attributed with the emergence of the PC as a home staple. Using commodity components, Intel processors, and Microsoft software usability has grown by leaps and bounds. Today we are seeing the advent of the under-$1000 computer. As a computing community, we do have a lot to owe to Microsoft.

However, it’s my opinion that Microsoft has not kept up with the rest of the computing community. Although their operating systems and applications have made computers easier to use in many cases, innovation has all but stopped, and the technology has failed to evolve along the paths it ought to.

Another of the most important technologies today is the telephone. Almost every day we use telephones to communicate with each other; we depend on it to spread news, enjoy friendships, and do business. We have come to transparently expect it to work: every time we pick up the phone, we expect to hear a dial tone. In similar ways we have come to take electric power, gasoline engines, and television broadcasts for granted. They simply work.

Similarly, much of computing is expected to “just work”, and it should be. There is very little about computers that are different than the above mentioned technologies. Indeed, computing is more complex, but the theory surrounding it is sufficient to make it work. And in practice, much of it does. Modern Unix operating systems and programs have progressed to the point of staggering reliability, capable of consistent performance over periods measured in years.

This is not the case with Microsoft.

Microsoft software, in my experience, is horribly unreliable. I spent almost two years administering a network using Microsoft NT Workstation and Server, their premier, industrial strength OS’s. Over the course of that time, I had to rebuild a number of systems after they mysteriously failed, both servers and workstations. My days were filled with seemingly random BSOD’s (“Blue Screen of Death”, a term coming from the white on blue text screen that Microsoft operating systems display when they crash) which interrupted my work, and halted the productivity of my company.

This was not my experience in the 3 preceding years dealing with a network that utilized both Novell Netware and several flavors of Unix (Solaris, AIX, SCO, and Linux). Systems stayed up for months or years on end, only being restarted after upgrades or patches (which increased even more the likelihood that a system would stay up longer). Thanks to the reliability of these systems, my job as a system administrator was to improve the functionality and services available to users: to help them do their jobs better. With NT, I spent much of my time “babysitting” systems and “putting out fires”. I could not be a “good” sysadmin with NT.

I am not the only one dealing with this. A friend of mine who is a respected consultant did some worth for a large corporation as they prepared to launch their Internet store-front. Having chosen NT found that they had to change to scope of their machine room attendant’s job to a full time, 24 hours post. This was necessary to have a person available at all times to reboot parts of their server farm as it failed. The truly sad part of this is that they accepted it: they didn’t see it as a problem that their server operating system was failing on such a regular basis! Would we accept such shoddy performance from our telephones, a 50/50 chance of getting a dial tone? I think not.

Today, after the release of Windows 2000, I can honestly say that Microsoft has not adequately attacked their reliability problems. Windows 2000 suffers from just as many reliability problems as NT4. I don’t have room or time to fully expose the breadth of my experiences here, but please believe me when I say they are vast.

After looking simply at the reliability aspect of computing, can we really say that Microsoft has kept up? I have to say, “No”. Programs much older than those from Microsoft prove to be far more reliable and usable. If we allow current trends to continue, we will have to lower our expectations even further to deal with the pains of software that has not kept up. As our society becomes more and more reliant on computing, as we have become reliant on telephones and electricity, we must demand greater reliability. I believe that we can have that level of reliability from places other than Microsoft.