Down for the count

One day the 2000 presidential election will be settled, but the problem that brought it on will not.

The inefficient voting process in several Florida counties, the one that has enriched the popular lexicon with ''hanging chad,'' is not an anomaly. Many other jurisdictions have cumbersome ballot mechanisms. Of the 2 million votes cast in Cook County, Ill., 120,503 ballots were thrown out because the voter failed to punch all the choices or punched two candidates for the same office, and that number is about par for past elections.

With current technology - optical scanners, for example - it is possible to devise a balloting and vote-counting mechanism that is foolproof, fraud-proof and fast. But the obstacles to doing so are nearly insurmountable.

The problem is that boards of elections are just not priority issues between elections. And because of costs and institutional inertia, there is great reluctance by state and local officials to change a system that ''has always worked.'' Broward, one of the disputed Florida counties, balked at spending $3 million on optical scanning voting machines. Riverside County, Calif., installed 4,200 touch-screen voting machines, but at a cost of $13 million. ATM-style voting machines run $3,000 to $5,000 each.

Nationally, there is no uniformity in voting methods. About 84 percent of jurisdictions use punch cards, optically scanned cards or the old-fashioned curtained booth with mechanical lever, around since 1892.

No great leap of technology is going to bail out the voting process. Everybody's universal panacea, the Internet, is too susceptible to hacker vandalism and vote manipulation; moreover, the Web and home and office PCs are not truly private. And the great advantage of a supervised common polling place is that the election can be seen as fair.

The states and localities should be free to conduct elections as they like, but the federal government has a compelling interest in the honesty and verifiability of federal elections. The new Congress - assuming some of those contests ever get settled - should direct the National Institute of Standards and Technology to develop technical guidelines for voting mechanisms that are, as said earlier, foolproof, fraud-proof and fast and that can generate quick and accurate recounts. It would be a legitimate use of federal funds to help jurisdictions willing to install these voting technologies if they cannot afford them on their own.

In a sense, research into voting machines is an insurance policy. Every election has its share of confusion and mix-ups, and it usually doesn't matter. This time it did.