 |
|
If the world really made sense, the businesses that are best and the brightest (and richest) got to where they are because the people in them work smarter than their competitors. But, of course, the world doesn't always make sense, and organizations at the top often get there in spite of their efforts, not because of them. Certainly that's the case with Microsoft, at least according to the picture Fred Moody paints in I Sing the Body Electronic: A Year with Microsoft on the Multimedia Frontier.
In extraordinary fashion, Moody was allowed virtually unfettered access to goings-on inside Microsoft from December, 1992 to approximately December, 1995, to chronicle the development of Explorapedia, a CD-ROM-based, interactive encyclopedia project. From the exuberance of preliminary planning meetings, to the anxiety of getting out the door a project that was late and over budget, Moody was there -- tape recorder in hand. The author was even allowed to tape "Bill Meetings" in which Microsoft chief Bill Gates alternately praised and pummeled "Sendak" (the codename for Explorapedia during the development cycle) team members.
If Moody's portrayal of the Sendak team is as honest as it seems, you have to think he didn't get many Christmas cards from Sendak participants once the book was published. Although just about everyone Moody encounters and describes is brilliant and talented, they also (for the most part) had little sense of or commitment to schedules, allowing the project to slip and slide its way out the door nearly a year behind schedule.
Multimedia projects at Microsoft were generally organized around two teams: content and development. Content specialists were, for the most part, designers and editors who wrote the words and created the pictures. Development teams were the programmers who built the systems. Central to Moody's book is the tension which gradually builds, as the content teams miss repeatedly key dates, change specs, and demand technically unfeasible features. All the time expecting that the programmers will make up the difference at the end.
At the same time, corporate politics are continually shifting back and forth in the background. From start to finish, content team members jockey for power, while freelancers fight for permanent positions (and the stock options that go with them) -- all the while with Chairman Bill lurking in the shadows. As if that wasn't enough, halfway through the project the Multimedia Publishing group itself fell victim to one of Microsoft's legendary reorganizations. It comes as no surprise that this leads to further bickering, with some team members wanting to go with one new division, while others want to go with different divisions.
Other tensions arise from top-down decisions that require programmers to use common toolsets and database engines being developed by yet other teams -- even though those tools are unfinished, unreliable, and unsuitable for the task at hand. Likewise, we get a peek at the legendary Microsoft hiring process, where job applicants are grilled by the programmers they will work for or with.
As an aside, and in retrospect, it is worth noting that I Sing the Body Electronic was written before Microsoft was an "Internet" company. In fact, to the best of my recollection, the words "Internet" or "World Wide Web" don't appear in the book. (In truth, this was kind of a relief.) However, throughout much of the book, you can replace "multimedia" or "CD-ROM" with "World Wide Web" or "Internet" and the sentiments being expressed are eerily similar to what we hear from Microsoft today. In nothing else, it is a testament to Microsoft that such a huge organization can be turned, so to speak, on a dime.
All in all, Moody paints a fascinating and (presumably) believable portrait of what it is like inside Microsoft. Still, it makes you respect even more the efforts of small companies that turn out great software without the luxury of meaningless schedules and unlimited budgets. In short, Moody makes you wonder how great Microsoft could be if they just put their minds to it.
-- Jon Erickson (jerickson@ddj.com)
There was a celebratory glitter in Kevin Gammill's eyes as he ran down the hallway toward [Carolyn] Bjerke's office. He paused outside her door and thrust his head in, shouting, "I'm at zero bugs on 'Encarta'!" Then he ran on.
A grand moment in liberation for a Microsoft developer, "zero bugs" is the end of months during which it seems that a project is doomed to remain forever on the verge of completion. After a project is built and run through tests, problems in performance -- bugs -- are entered in a file called "Raid" and assigned to specific developers to fix. Bug-fixing can be extremely tedious work -- a whole day can be devoted to fixing one or two -- and "Encarta," with its 2,100 bugs, took Gammill and his cohort, Jay Gibson, months to clean up. Each bug required a search through "Encarta's" code, looking for the deletion of an error, or the rearrangement of existing code. Occasionally, newly written or newly arranged code would introduce new bugs, requiring yet another round of repair.
Now, with Gammill at zero bugs and Gibson only a few away, the March 8 ship date for "Encarta," only a week off, looked secure, and Gammill could look ahead for the first time in years at a long stretch of relative relaxation. Before he could begin work on "Sendak," its design would have to be complete -- a circumstance that was two or three months away, to judge from the behavior of this teammates...
It was not only "Encarta" that had kept Gammill from closer involvement in "Sendak's" planning. Gammill had no patience for the often directionless path that product design took, with groups of people randomly wandering all over the conceptual map. He preferred being given specific features to evaluate, thinking about them on his own, then coming back to designers with a yes, a no, or an assessment of how long it would take him to implement their design. He detested trying to explain development problems to nondevelopers.
-- Page 113
| Readability |
|
| Originality |
|
| Organization |
|
| Accuracy |
|
| Consistency |
|
| Depth |
|
| Timeliness |
|
| Editing |
|
| Overall Value |
|
Explanation of ERCB rating scale: No stars = unacceptable, 1 Star = marginal, 2 Stars = average, 3 Stars = above average, 4 Stars = exceptional.