E-mail in bondage.

It's been one of those weeks. All the computers I have passwords for decided at the same time to deny me the ability to go online. No e-mail. No research capability. No chance to touch base with my cyberspace colleagues.

Why? Beats me. If I knew, I'd fix it without as much effort as I've put in on this. I've spent hours talking to countless technology support reps on hotlines and more hours with technicians hovering over my shoulders watching the strange things computers can do.

I've messed with winsocks and modem drivers, maps and preference files, LANs and PPPs. I've erased hard drives, then loaded and reloaded software in a variety of versions. And what have I concluded?

Computers still don't work the way they should.

No revelation there, of course. That ol' term "user-friendly" has been worn out from all our requests, yet computers still aren't. I say this with full knowledge that I understand computer logic (oxymoron acknowledged) better than most people. And I'm still going to have to provide this "Digital Knowledge" column in hard copy to the magazine editors.

Someone who really knows computers, Danny Hillis, said in the August 11, 1997, Newsweek: "You don't think of a pencil as technology, because it just does what it's supposed to do. You don't think of a book as technology because it's been refined enough that it becomes invisible.... You still think about computers a lot, because they're so bad."

And Hillis, founder of Thinking Machines supercomputer company, goes on to say: "I think technology is all the stuff that doesn't work yet."

Some people, like me, vent in frustration when computers don't work as well as they should. Other people, like Jaron Lanier, invent new computers when things don't work as well as they should. Lanier, who coined the term "virtual reality" a decade ago, currently is lead scientist for the Tele-Emersion Initiative trying to make virtual reality work across long distances over networks such as the Internet.

Lanier tells of his school days when math professors gave out difficult problems. Students complained and vented in frustration. "Then, one day you understand. And once you get it, you realize it's really simple. It's simple once you get it.

"It's hard to understand, but easy to know.

"That led me to think that the way we express math must make it seem a lot harder than it is. We must be using the wrong language," he said.

"It led me to believe that the degree to which a human can understand something that's hard to understand is largely determined by how that thing is presented."

For business communicators, that's not such a novel concept - and we probably didn't have to suffer through logarithms to arrive at the conclusion. Having interviewed candidates for several job openings in the last year, I can attest to the fact that the statement "I want to work in public relations because I like working with people" has been replaced by "I can take difficult or complicated information and write it so that people understand" as the most common interview response.

Still, the fact that people who shape our computing experiences recognize that these machines don't work well, that current computer interfaces are a lot more complicated than necessary, should be encouraging. Nobody's resting on laurels, though the current way computers present us information is a vast improvement over 10 years ago.

Yet, there's no reason computers can't be smart enough to think through the processes necessary to find glitches in init strings, confirm software compatibility, and let my e-mail go. Why can't I type in "Fix my online connection" and have the computer itself analyze and solve the problem - or at least report back to me and I'll initiate the fix?

In fact, according to Lanier, making computers smarter is easy work - relatively speaking, of course.

"There are lots of people who make machines smarter. If you come up with new algorithms to analyze a big data base to notice patterns you wouldn't have noticed before, that's making machines smarter.

"That's important work. But, if you can't communicate the results of that clearly to the humans involved, it's wasted work," he said. "In every situation I've ever been involved with, making the computer clearer actually yields more productivity. It's the more important process."

In Lanier's mind, "virtual reality is the end point of trying to make computers clearer." He's speaking of his work with military simulations and remote equipment testing at large industrial companies, for example,' though you may think of virtual reality as those goggles and headgear for fun and games. ".Simulation of alternative reality may be fun, but the real value of virtual reality is when it is used collaboratively," he said.

Imagine sharing an alternate world with a coworker a thousand miles away.

I'm just trying to imagine sending text to cyberspace colleagues oh a computer that's working right. But I'm optimistic. These intelligent computer scientists will create impressive, pricey systems, parts of which will trickle down to my level. In the meantime, I'll be back online in a couple of days, I'm sure. I hope.

Sheri Rosen, ABC, is director of employee communication at USA.A, a Fortune 200 financial services company. She invites your thoughts at [e-mail]