http://www.blackboxvoting.org/?q=node/view/78
another article: http://www.cbsnews.com/stories/2004/07/28/sunday/main632436.shtml
combine this with the lack of a paper trail in the voting machines themselves (the above is about the central tabulator that counts the votes) and the Diebold CEO's declaration that he is "committed to helping Ohio deliver its electoral votes to the president next year" (the reference is within the article, which is worthwhile to read on its own) and we have an election that very may well be rigged on our hands...The central tabulator is far more vulnerable than the touch screen terminals. Think about it: If you were going to tamper with an election, would you rather tamper with 4,500 individual voting machines, or with just one machine, the central tabulator which receives votes from all the machines? Of course, the central tabulator is the most desirable target.
Findings: The GEMS central tabulator program is incorrectly designed and highly vulnerable to fraud. Election results can be changed in a matter of seconds. Part of the program we examined appears to be designed with election tampering in mind. We have also learned that election officials maintain inadequate controls over access to the central tabulator. We need to beef up procedures to mitigate risks.
Much of this information, originally published on July 8, 2003, has since been corroborated by formal studies (RABA) and by Diebold's own internal memos written by its programmers.
another article: http://www.cbsnews.com/stories/2004/07/28/sunday/main632436.shtml
"The concern that I have is not that somebody will tamper with the machine on Election Day and change the outcome. The concern I have is that those machines will be programmed from the start to favor one candidate over another and not to actually record and count the votes," says [ed: Avi, a CS prof at Johns Hopkins] Rubin.