InformationWeek

May 17, 1999

Windows 2000 Beta 3: Bang The Code Aggressively

By Stuart J. Johnston

s promised, Microsoft delivered the third and final beta version of Windows 2000 at the end of last month, meaning that users will begin receiving it in coming weeks. Company officials say they still think they will be able to ship Windows 2000 by year's end.

Microsoft officials say this version constitutes a feature-complete release, including plug-and-play support for as many as 7,000 different hardware devices and at least 150 different notebook computers. It also provides Windows Terminal Services support as well as the Windows installer technologies that enforce tighter restrictions on how applications can be installed.

The business preview program discs come with a license to install five copies of the desktop version -- Windows 2000 Professional -- and two copies of Windows 2000 Server. Therefore, with approximately 400,000 copies of the preview program discs expected to ship, it is likely that more than 1 million users will have at least some opportunity to hammer on beta three.

Let's all hope that a lot of people hammer on it hard, and if it's not soup yet, that they tell Microsoft loud and clear not to ship it until it is completely cooked. Microsoft officials insist they will not release Windows 2000 until users tell them it's rock solid. Granted, having the largest beta-test distribution in history will certainly help in that regard. There are so many different hardware configurations out there today that it is hard to predict what will work and what won't, and even with 10,000 machines testing it every night inside Microsoft itself, there still may be configurations that no one thought of testing.

So, certainly it is laudable that Microsoft is trying to get a very large segment of the user base testing the operating system before it actually ships -- instead of the other way around, something for which industry wags have previously criticized the company. At the same time, however, I wonder if there isn't a little of the "Mythical Man-Month" problem that may be creeping into this process.

For anyone who has forgotten or who is too young to remember the analogy, there was a well-regarded software-management text titled "The Mythical Man-Month: Essays on Software Engineering," by Frederick P. Brooks, published back in 1975, and reissued in paperback just four years ago.

One of the extrapolations of the Mythical Man-Month that was frequently bandied around back in the mid-1980s when I worked in large-scale development organizations was the joke that if it took one man a month to dig a big hole, then management would automatically assume that 30 people could dig that same hole in a day. It certainly seemed to us worker bees that this was management's attitude, and I can't help but think that oblivious management thinking like this spurred the popularity of Scott Adams' Dilbert cartoon.

So, while holding the largest "official" beta test in history may be a good idea, I have to wonder if the concept doesn't hold the seeds of some potentially serious problems. All software has bugs. Windows 3.1 shipped with more than 500 known bugs, as a confidential Microsoft report that I obtained at the time demonstrated.

But history teaches that bugs in new systems tend to get worked out over time -- after all, what are "service packs" and "patches" for, if not to fix things that are discovered and repaired after the product ships? It's a process that has been compared to aging fine red wine.

With that many users hammering on Windows 2000 beta three, the really big bugs will likely get identified and, one would hope, fixed by the time of the final release. But there is a standing joke and truism, often repeated by Microsoft product managers and other senior officials, that "shipping is a feature," too.

That means that at some point, Microsoft has to decide Windows 2000 is "finished" and ship it. Right now, it's still telling us that the end of the year is the target.

Additionally, programmers talk about the "ripple effect" in fixing bugs. Software programming is still more of an art than a science, and every engineer has his or her own ideas about how to solve any given problem. A piece of code may appear to work in most circumstances, until it is tested under a particular stress mode that may cause it to fail or behave differently than expected.

While it might seem logical that fixing bugs should be straightforward once they are discovered, what often happens is that fixing one bug exposes another in a long cascade of problems that need to be fixed one after another -- thus the term "ripple effect."

With roughly 29 million lines of code in the desktop version of Windows 2000, and taking into account the ripple effect, I still wonder whether the pressure to ship something this year may eventually threaten to short-circuit the need to make sure that every single thing is fixed before Windows 2000 goes out the door.

Microsoft officials say they clearly understand the need for rock-solid code. Indeed, InformationWeek Research surveys of IT decision makers have consistently found that above all else, Windows 2000 must be entirely reliable and robust in order to be taken seriously. That being the case, I hope all the beta testers hammer on that code as hard as they can. And if they still have problems with the code, no matter how close it gets to the end of the year, testers need to make sure that Microsoft hears their feedback. Company executives say that is what they want -- honest feedback.

Company stockholders may not be happy if Windows 2000, and the huge revenue boom it will likely spur, is held up yet again. But 100 million users will ultimately appreciate that honest feedback when it's time to roll out Windows 2000.

AuthorITies Archive

Send Us Your Feedback

Top of the Page


Karyl Scott: Enterprise View Karyl will explore the business and technology issues surrounding enterprise systems.	Charles Pelon: Eye On IT Charles explores IT management issues and strategies that business and technology managers face.	Jason Levitt: Internet Zone Jason focuses on the strange, egregious, and the standard technologies of the intranet/Internet.	Sean Gallagher: The Bleeding Edge From his vantage point of managing editor of InformationWeeek Labs, Sean will explore the impact of new technologies on the evolving world of electronic business.