S.C. election officials insist the machines voters will use Nov. 2 will give an accurate tally, but a growing number of people are less sure.

For six years, the state's voters have used the iVotronic touch-screen system made by Election Systems & Software, a Nebraska-based company that's one of the major players in the industry.

The state bought these 12,000 machines with federal dollars doled out in response to Florida's "hanging chad" controversy during the 2000 presidential race. They replaced a hodgepodge of voting systems used by the state's 46 counties.

Despite criticism of the iVotronic system in other states, the reliability of South Carolina's machines hadn't been much of an issue until Charleston County Councilman Vic Rawl questioned them after what many considered his shocking Senate primary loss to unknown Manning resident Alvin Greene.

But Rawl's loss -- and the puzzling discovery that Greene won big on many counties' machines while simultaneously losing to Rawl in paper absentee ballots in those same counties -- breathed fire into a debate that had smoldered quietly among software experts, private citizens and some members of the S.C. League of Women Voters.

"We put up pretty substantial circumstantial evidence that it couldn't have happened but for the machines," Rawl said of his loss.

The issue is by no means unique to South Carolina. A recent report from the Brennan Center in New York analyzed problems with voting machines in other states and said more should be done to document their reliability.

"Unlike makers of other commercial products, voting machine manufacturers are not obligated to report malfunctions to any government agency," the report concluded, "and election officials and the public are often totally reliant on the private companies that sell and service the equipment to voluntarily keep them aware of potential problems."

While everyone agrees that concerns about the machines shouldn't keep anyone from casting their ballot next month, they also know this is a prime time to raise the issue, as voting -- and how it works -- moves to the forefront of people's minds.

Still, the voting machine debate seems stuck in a Catch-22. Supporters say there's no evidence that the machines have caused problems; detractors say the fact that the machines can't be double-checked for errors is the problem.

Further complicating the debate is its complexity: Election officials admit there's no perfect election technology that guarantees accurate and secure results.

There's another Catch-22, too, when it comes to their security. The machine's software is held in secret by ES&S for competitive reasons, while the state's security plan can remain secret under the state's Freedom of Information Act.

"We wouldn't want to give a thief a map of the schematics of our burglar alarm system," State Election Commission spokesman Chris Whitmire said.

While the iVotronic machines haven't had a perfect run in South Carolina -- several shut down in Horry County in 2008 -- there had been little talk of their affecting an outcome until June 9. That's when Rawl and his supporters began examining how he lost the Democratic Senate primary by 20 points to a man who didn't campaign (and who later was discovered to be facing a felony charge).

Truett Nettles, a former Charleston County Board of Elections chairman and the lawyer who handled Rawl's appeal, said his best guess about what happened is that every fourth vote for Rawl somehow was switched to a vote for Greene on the state's voting machines.

"The result here was an anomaly. It wasn't just a freak thing. It was statistically impossible," Nettles said.

However, the state party rejected the appeal because Nettles and Rawl could offer only circumstantial proof that it happened. And Nettles said that's the problem.

"We've got to have a system capable of going back and reviewing it and verifying it," he said.

The number of people questioning the machines seems to be on the rise. Frank Heindel, a Charleston businessman, began asking about the machines after the June primaries and retrieved reams of records showing assorted problems -- problems election officials insist didn't affect any results.

"I have a hard time being confident with it when you have a secret security plan and a secret certification report," he said.

"Nobody is paying attention. Nobody is looking over anybody's shoulder."

Duncan A. Buell, a professor with the Department of Computer Science and Engineering at the University of South Carolina, has studied the machines along with the state's League of Women Voters.

Buell appeared before state election commissioners late last month to discuss a 2007 report from Ohio that found the machines can't be reliably used for elections and that no procedures will reduce the risks.

But he shares the same concern about software glitches and proprietary software and also finds Rawl's loss to Greene the only anomalous event in the state to date.

"A lot of people have said we've had no problems, but no one can say we've had no problems because no one knows what the truth really was," he said.

S.C. League of Women Voters President Barbara Zia said she shares Buell's concerns and finds it unfortunate that the issue isn't on anyone's radar screen until there's an election.

Meanwhile, the voting machine issue will continue to percolate nationally. A Libertarian candidate in Georgia's secretary of state race is making doubts about his state's voting machines a keystone in his campaign.

And the federal government is expected to set new standards for voting machines and decide what to do about the Brennan report's advice to: create a publicly available, searchable database of voting system glitches; to mandate rules for when voting manufacturers must report to the database; and to give a federal agency power to ensure vendors report to that database and take steps to fix any problems found.

"Bottom line: lost votes damage public confidence in the electoral system," said Lawrence Norden, the attorney at the Brennan Center for Justice who authored its recent report.