Matt Blaze is right on point, as usual; in fact, he probably does not go far enough in his critique of scientific publishing. Commercial publishers from Nature to those on the long tail are as bad or worse. The "in your face" commercial paywall in front of most technical journals is a frequent user experience here on Hacker News. And counterproductive to doing good science and engineering.

If the free exchange of information promotes the growth of human knowledge, then free it ought to be. The ACM an IEEE digital libraries have improved and simplified research in many ways. The fact that they are behind a paywall and available only to members (and some institutions) is a negative.

Many alternative venues for publication are appearing, for example, Arxiv, http://arxiv.org/. Google's Knol service is also a reasonable place for articles about new work. But they do not have the cachet of a peer-reviewed archival journal.

ACM and IEEE are membership organizations. Perhaps we need a grass roots effort to change the copyright policy to, say, one of the creative commons licenses and open the digital libraries and electronic journals to all comers.

On the other hand, a well run, professionally edited journal costs money to produce. In a world where content is free, how should we monetize publications so that the staff gets paid?