Like any writer with a healthy ego, I'd love to push my way into your computer.

I've written dozens of articles over the years, covering technological advances ranging from fiber optics (late seventies) to gallium-arsenide semiconductors (mid-eighties) to the World Wide Web (the nineties and forever). I've poured hundreds of thousands of words into printed magazines that have then made their way through rain, snow, and the gloom of budget cuts into the offices and homes of people whose desks and shelves sport great stacks of unread journals, conference proceedings, books, and other printed information. My words fell silently upon these massive piles, and as a writer I could only hope that readers would someday be moved to thumb their way to my byline, before another delivery of mail buried my lucid explanations - my descriptions bursting with clarity - for good.

The problem, obviously, is that I counted on the magazines' mere existence to stimulate a subscriber to open and read them - in particular, to read my pieces. I now see how naive I was. What I should have done was to get on the phone and call the hundred thousand people that might have some possible interest in my latest report from the high-technology front. I should have instructed each person to retrieve the magazine from their mailbox, open to the correct page, and start reading. That's the kind of proactive, assertive behavior that gets one's signal noticed amid the squawk of info-noise that fills our lives.

This is, in effect, what is offered by a scientific journal alerting service, which e-mails to its clients bulletins announcing new papers that match a preselected range of topics. An oncological researcher might, for example, specify an interest in studies of neuroblastomas. Publication of any such article triggers an e-mail to the oncologist, giving a citation and abstract. The oncologist can then either hoof it over to the library to read the article or, for an additional fee, request a copy of the full text, which arrives the old-fashioned way - through the mail, as hard copy - within a few days.

Pushing hardest on such "push" techniques is the Institute for Scientific Information (ISI), which has been offering the Journal Tracker e-mail alert service since April 1996. For $99 a year, ISI provides tables of contents for as many as 25 publications selected by the client. A more customized service, for $395 a year, gives individuals the ability to fish around in all 16,000 journals and conference proceedings of which ISI keeps track. This service avoids the tunnel vision that can result from always drawing information from the same few sources. One user of this service, statistician Roy Robertson of Rohm and Haas in Bristol, Pennsylvania, explains the value of this breadth. Robertson says he is constantly scouting for new and more reliable statistical modeling methods to help predict the properties of the materials his company develops. Sometimes, he says, new methods crop up not in the statistics journals but in publications devoted to other fields, such as chemical engineering. Articles appear too seldom to make regularly reading these journals worthwhile, but he finds the occasional insight from other fields highly valuable.

But to some advocates of electronic publishing, e-mail alert services are merely holding actions that temporarily give traditional journal publishers something to do. With the rise of electronic communications, scholars have less need for traditional publishers, says Andrew Odlyzko, head of mathematical and cryptographic research at AT&T Bell Laboratories. "What publishers do is superfluous," Odlyzko contends. Researchers can handle the distribution of new knowledge themselves through academic editors and peer reviewers, who work on a volunteer basis in any case. "For publishers to survive," Odlyzko says, "they will need to develop enhanced services" such as ISI's offerings.

So far, scientists have not exactly embraced the idea of e-mail alerts. Academic researchers already tend to have a finger on the pulse of their niche. The bigger market is among industrial scientists, who are often eclectic and who spend their R&D time mostly on D. All told, ISI counts only a few hundred subscribers for its service, which has been offered since April 1996. That's a pretty small slice of the research community.

One reason may be that this is a hybrid, half step toward the kind of full electronic dissemination that information-hungry knowledge workers crave. Once people start "living" on the Internet, any process requiring them to detour through the physical world can seem an intolerable inconvenience. Thus Robertson of Rohm and Haas hesitates not at all when asked what would improve the service. "I want to link to the paper on the Web."

But that's not going to happen anytime soon. With few exceptions, journal publishers - afraid of losing subscription revenue - do not make their full-text electronic files available to ISI, which therefore cannot pass the files along to its clients. Copyright concerns continue to cloud the future of electronic distribution, and the scientific societies that publish many journals are not sure which way to turn. The big question, says Robert Park of the American Physical Society, is: "How do you get paid?"

Scientists on the front lines of research, however, probably care less about the economics of journal publishing than about the ease with which they can ferret out the information they need to get their jobs done. E-mail alert isn't the final word, but it's a sensible start.

"The Slow Evolution of Electronic Publishing" - Andrew Odlyzko's paper on the history of online publishing and the future of print publishers. His site also contains several other works on electronic publishing and electronic commerce and on research, technology, and society.

"Push! Kiss Your Browser Goodbye: The Radical Future of Media Beyond the Web" - active technology makes the Internet grow in size and complexity, and the tools to handle it will transcend not only browsers, but computers. An in-depth study from the March 1997 issue of Wired.

Software Agents Group - an MIT Media Laboratory research team developing new techniques and algorithms relating to software agents, the "proactive, adaptive, personalized, and autonomous" systems that aid users of applications.

"Agents Are Getting to Know All about You" - active agents are guided by, not directed by, the preferences of the people they survey. From the November 1997 issue of Wired.

Information Filtering & Collaborative Filtering - a slide-show-style seminar about agent technology, with linked examples of resources, comparative tables, and feedback from students.