I've seen an astonishing number of posts (here, here, and here, among others) claiming that evidence exists that the presidential election returns were tampered with, hacked, or otherwise modified to give Bush a victory. Specifically, people are pointing to two things: First, one precinct in Ohio mistakenly reported more votes for Bush than voters who showed up at the polls; and second, a few small Florida counties with large Democratic majorities voted overwhelmingly for Bush.
Neither of these is the slightest indicator of fraud, and neither has anything to do with electronic voting. The problem with the vote in that Ohio precinct is unfortunately common in initial reported returns. In fact, I'd guess that just about every state in every election has at least one precinct that reports results incorrectly. This happens all the time. It's nearly always human error at some step -- someone wrote down the wrong number, misheard results reported over the phone, or something like that. In this year's Pennsylvania Senate primary, one precinct in Bucks County initially reported about 100,000 votes for Arlen Specter. That mistake was so noticeable -- it swung the incoming Election Night returns from a close race to a large margin for Specter -- that it was corrected within an hour or two. Smaller mistakes like the Ohio one are always caught in a recount because the first step in a recount is to go back and double-check the totals on the actual machines. (Though popular perception would have you think that most recounts involve ballot analysis like the hanging chad problem in Florida in 2000, that's rarely the case.) Even in races that don't have recounts, they're typically caught when the official returns are reported. That's what will happen in Ohio.
As for the Florida counties, there's nothing suspicious here, either. A quick glance at the 2000 election returns in the same counties shows that they went overwhelmingly for Bush then, too. The ones in question voted slightly more Republican in 2004 than they did in 2000, but that's not surprising at all given the overall election results. Bush did better across the board in Florida in 2004. In fact, those counties' results would be far more suspicious if they voted for Kerry, or even if the two candidates tied.
None of this is to say that electronic voting is fraud-proof or that paper receipts shouldn't be mandatory. I write software for a living, so I will never believe any electronic voting company's claims that they can be certain their software is bug-free or tamper-proof. Only voter-verified paper trails will protect against that. But though the system may be insecure, there's simply no evidence of fraud this time around. Those of us who came out on the losing side would be far better off spending our time trying to figure out how we can move people into our column in 2006 and 2008 than trying to run after electronic voting fraud as if we're hunting for weapons of mass destruction in Iraq.