There was a small news item in last month's issue of this magazine. The Business of Government section reported on a new online program in Missouri that gathers disease data from 50 labs and hospitals and tells the Health Department almost instantly if something resembling an epidemic is loose in the state. Health Department employees used to have to call those hospitals regularly to find out what was going on. Now they just know.

Things like this are happening all over the place these days in American government. Vital information that used to be collected slowly, expensively and unreliably is now available without much effort 24 hours a day. Lives are saved as a result.

I've always been skeptical about dramatic claims of governmental breakthrough, and I don't think I'm a sucker for them now. But even a registered skeptic can't help noticing all the ways information is making government function much better--at every level of the system.

In the same issue as the Missouri item, we ran a story about the way Philadelphia dealt with a snow emergency earlier this year. The city desperately needed to clear its storm drains, but there was no way workers could get to all of them. So they asked for citizen volunteers. The drains were invisible under two feet of snow, but it was easy to find out exactly where they were, because each drain was pinpointed on an electronic map available to anyone with a computer.

These are nice stories. Technology is starting to make government look good. Meanwhile, however, politics is making it look terrible.

The same month that I learned about disease monitoring in Missouri and drain clearing in Philadelphia, I was also pondering a collection of political horror stories sufficient to discourage even the most determined optimist. California was conducting a clownish and fundamentally illegitimate recall campaign against a governor whose only real offense was mediocrity. The Texas legislature was paralyzed by efforts to redraw congressional districts for no purpose other than blatant partisan gain. Alabama voters were repudiating a tax reform program that would have transferred money to most of them--including most of those who voted against it.

It's hard to think about all of this without coming to a stark conclusion: The sharpest division in American civic life these days isn't between factions, or ideologies, or between competence and incompetence. It's between government, which is getting better, and politics, which is getting worse.

It may be arguable whether medical data-sharing and geographic information systems are creating profound changes in everyday life. But other governmental innovations made possible by technology clearly are. When you had an emergency at your house three decades ago, you rummaged around for the phone number of the police department and hoped you were picking the right extension. The introduction of 911 has been an immense improvement. The COMPSTAT system of tracking crime patterns electronically--rather than just responding to incidents as they occur--is at the heart of the decline in violent crime that took place in almost every major American city in the 1990s.

Information technology is not only refashioning government, it is breeding a generation of bureaucrats who are irrepressibly enthusiastic about its use and impatient to move to new levels of sophistication. They are at conferences and on the Internet, comparing notes about the newest techniques in performance measurement, "smart" transportation, homeland defense and on and on. In this way, at least, the first decade of the new century is beginning to look a little like the Progressive Era in the years before World War I, a time when idealistic technocrats and civil servants argued that scientific management promised a bright new era of governmental efficiency.

It's no mystery why that kind of enthusiasm exists. Quite a few of the problems of government are ultimately problems of information. We're slow to recognize epidemics because we don't know who's getting sick at the other end of the state. We fail to prevent homicides because we don't have precise data on when and where they are occurring. These are problems against which breakthroughs in information technology can work wonders. And in the past decade, our sophistication in information technology has grown by multiples almost too large for most of us even to grasp.

Wherever problems in the public realm are largely the consequence of scanty or inaccurate information, we are making genuine progress in solving them. That's why government is getting better. But where information isn't the problem, it can't be the solution, either. That's why politics is getting worse.

Unfortunately, the more skillful we become at gathering and massaging data, the more we are tempted to apply the same techniques to problems that information can't really solve. The technocrats of the Progressive Era thought that if they could perform enough time-and- motion studies on the public workforce, they could usher in a revolution in worker productivity. That turned out not to be true. The architects of the No Child Left Behind Act believe that if they can compile enough data on school performance, they can prevent students from failing. That won't turn out to be true, either.

So one of the unpleasant side effects of the current boom in public technology is an inappropriate and often naive extension to maladies it can't cure. But there's an even more sinister side effect. It's the capacity of technological innovation, when unaccompanied by societal restraint, to make the political system even worse than it was before.

Think about some of the grossest excesses of this very gross political year. Drive-by gerrymanders, for example, unrelated to the regular reapportionment that follows the decennial census. These used to be extremely rare, and for a couple of very good reasons. First, new district maps were troublesome to produce. Second, the statistics on voting behavior were problematic. Partisan mapmakers could easily scheme to draw two seats for their party in a legislature, and find out later they had cost the party two seats.

Computers and geographic information systems have changed all that. It's still impossible to predict the future voting habits of any constituency, but it is possible to draw districts on a map with considerably greater confidence about how they will behave in the real world. That's a temptation to mischief, and it's a temptation that ambitious politicians all over the country are increasingly unable to resist.

Then there is the link between technology and negative campaigning. Televised campaign slurs have been part of American politics for half a century, but until the 1980s, they were held at least partially in check by the time required to produce and distribute them. It could take several days to create an ad, several more to respond to it, then a few more to respond to the response. This in itself didn't make politics any more sportsmanlike, but it made the pace slower.

Electronic tape editing cut the response time almost to zero. If I want to, I can slander you tonight, and you can slander me by tomorrow morning. This increases the pacing and intensity of the whole enterprise, and gives campaigns the quality of an escalating arms race in which the immediate use of almost any lethal weapon is justified. It's hard to see how anybody benefits from this.

Finally, there is polling. It would be foolish to argue that trimming one's convictions is a weakness that old-time politicians never indulged in. Nevertheless, old-timers with a weakness for pandering-- even the smartest among them--couldn't always be sure how large a majority they were pandering to. In the absence of modern polling techniques and data-crunching technology, it wasn't easy to know where public opinion stood on every issue every day of the year.

In general, this was a good thing. I sometimes find myself thinking back to the congressional elections of 1970. The overriding issue was the Vietnam War. For the most part, candidates either supported it or opposed it. Many of them would have taken the same position regardless of what the voters thought, but I can't help thinking that the inability to gauge public sentiment in precise detail helped free them to say what they really thought. Maybe they were on the side of a 53 percent majority, or maybe it was a 47 percent minority. Nobody knew.

Contrast this to the gyrations of Democratic presidential candidates and congressional leaders over the past year on the subject of the Iraq war. Most of them have shown an impressive ability to match their beliefs to the cross-tabulated poll numbers of the moment. It remained for the octogenarian U.S. Senator Robert Byrd, a pre-modern politician if there ever was one, to remind everyone what taking a stand and sticking to it used to be like. The unpleasant truth is that in modern politics, technology and courage are often inversely related.

Thus the paradox: Information is remaking government for the better, at every level from city hall to the federal civil service. At the same time, it is helping to drain the political process of fairness and integrity. This is not because technology is a bad thing, but because the most important problems of politics are beyond its reach.

The state of American politics will improve when it reclaims a measure of public trust, so that leaders are allowed to lead--not undercut by spurious recalls and initiatives and trash-talk campaigning. I have no idea when--or if--this will happen. I know it's something that machines and data are powerless to fix.

In the short run, perhaps we should take some satisfaction in the growing number of public problems that machines really can help us with: tracking epidemics, responding to emergency calls, pointing the way to city drains that citizens can help unclog. I don't know about you, but I'm grateful for small favors.