The Community for Technology Leaders

Productive and healthy debate

David H. , IEEE Micro

Pages: p. 6

This issue of IEEE Micro is a little bit different, to say the least. It includes four debates from the Computer Architecture Research Directions (CARD) workshop that was held as part of the 13th Annual IEEE/ACM International Symposium on High-Performance Computer Architecture in February 2007. The organizers approached me about a special Micro issue about CARD and I said, "Why not?" We need to take risks with new format ideas to avoid stagnation. Moreover, a debate brings into focus the important issues of a particular topic, and it's a great way to extract position statements from experts in the field. Plus, debates are a heck of a lot of fun—conference attendees flock to panel sessions hoping for controversy, name-calling (good-natured, of course), fisticuffs, whatever.

In a society whose attention span seems to be dropping at an exponential rate, and in which soundbites or just plain obnoxiousness attract far too much attention, our scientific communities are a refuge of productive and healthy debate. A productive, healthy debate is informative and insightful, and collectively moves the debaters and the audience closer to a potential solution or agreement by examining the issues from multiple viewpoints and critiquing those arguments. Ideally, the debate is not about the debaters themselves—who is better than whom—but focuses on the issues. Good debaters make convincing arguments against the opposite viewpoint, while avoiding blatant, unnecessary attacks on character. In fact, an effective debater—one whose arguments resonate with the audience—will subtly boost the character of the opponent while arguing against the position. By contrast, an unproductive debate degrades to unhealthy personal attacks that have no basis in supporting the substance of the arguments. The objective becomes personal power over the common good.

Thankfully, productive and healthy debate abounds in the computer architecture community. Although I have at times witnessed, and cringed at, an unhealthy flavor, it's relatively rare. When I entered this community at ISCA 1997 as a first-year assistant professor—I still remember the elevator doors opening to a room full of strangers—I quickly felt welcome. Eventually that welcome extended from the newest members of the community to the preeminent computer architects who are the most influential in the group. Those at the top set the tone, and we are extremely fortunate to have leadership that sets a good example for our community: They debate using facts, knowledge, and experience—and not personal power. (Okay, perhaps I'm getting a bit utopian, but I stand by my general viewpoint.)

I think you'll find the debates captured from the CARD workshop to be quite informative—and healthy. Whereas many of our workshops resemble "mini-conferences" (with WDDD, the Annual Workshop on Duplicating, Deconstructing, and Debunking, perhaps the most notable exception), the CARD format of "mini-debates" is a refreshing change. Leading experts in each of the areas take opposing viewpoints to expose differences of opinion that can be critiqued. Here you'll find the civil, informative, at times entertaining debate of the issues that makes our community a place we should not take for granted.

I wholeheartedly thank Josh Yi, Derek Chiou, and Resit Sendag for the tremendous effort they put into this issue. They conceived the idea, performed the initial transcription of the articles, got the moderators and debaters involved, rounded up illustrations, and worked with the IEEE Micro staff on revisions. If you enjoy what you read here, send them an e-mail of thanks.

I always welcome your feedback at

57 ms
(Ver 3.x)