University of Cincinnati professor George Bishop laughs at the prospect of a definitive and accurate presidential vote count in Florida - or anywhere else.

Random measurement error - a statistical reality that afflicts nearly every process involving large numbers - will bar the discovery of true, final numbers, he says.

''I hope no one believes there is one total, absolute count here. Nobody's ever going to know. Americans think there must be a real number, some true number. Guess what? There isn't,'' says Bishop, a political scientist.

Scientists know random error creeps into research, so they will run calculations several times and choose a number somewhere in the middle as representative of the true answer.

''Maybe we'll approximate or get close, but who knows what the real number is. Every time you count those ballots there is measurement error involved - the random error of counting machines and processes,'' Bishop says.

And, recounts only exacerbate the accuracy of numbers because you introduce systematic error - human judgment, for example, on whether a ballot chad is dimpled - that typically leans in favor of one candidate, Bishop says.

Random measurement error, however, does not favor one candidate over another. ''You could run those punch cards 100 times through machines and get a different number every time. But, in the end, it would basically be a wash,'' Bishop says.

Kenneth Arrow's Nobel Prize-winning ''Impossibility Theorem'' established there's no such thing as a perfect voting system in races involving more than two candidates.

But, mathematicians deem the plurality voting system - in which the candidate with the most votes wins even without capturing a majority - a flawed measure of candidate support.

According to the Center for Voting and Democracy, plurality voting only measures the amount of focused, core support a candidate receives, with his or her breadth of support of no consequence.

There are other voting systems:

Condorcet measures breadth of support among voters, but disregards strength of support. A winner may not be the top choice of any voters, but would compare favorably one-on-one with other candidates.

The Borda Count - similar to the proportional representation system once used to elect Cincinnati City Council - asks voters to rank their preferences.

Approval voting allows voters to cast ballots for as many candidates as they wish, with the candidate deemed the most acceptable winning election. However, it fails to distinguish between first choices and those with a weaker level of approval.

Instant runoff voting requires strong enough support to avoid elimination as voters rank candidates in order of choice, but still requires a candidate to receive broad support and capture a majority of votes.

Candidates are progressively eliminated based on the number of first-place votes they receive. As candidates are dropped through recounts, their supporters' second and third-place votes, and so on, are cast until a winner is established with a majority of votes.