We KNOW that paperless DREs are less verifiable than Op Scans. This is one reason not to rely too heavily on the exit polls as they did not always deviate according to machine type. But there are reasons why this may be true and past performance is no guarantee of future results anyway. Any system can be hacked, or not, for any given election!
See these links on the Dixiecrat counties:
<
http://residentbush.com/Aftermath-2004_Florida-Results.htm><
http://ustogether.org/election04/mitteldorf/Liddle.htm>2000 was not much different. Same trend to a smaller extent. Somewhere on the ustogether site, there's a comparison between 2000 and 2004 but it's hard to find now.
Both Op Scans and DREs can be hacked! OBVIOUSLY, Op Scans are better than PAPERLESS DREs. You can't verify the vote at all on paperless DREs. If they want to debate DREs w/VVPARs, vs. Op Scans, that's another story. So see this page:
<
http://www.nyvv.org/paperballotVsDRE.htm>Op Scans are still better!
In general precinct count Op Scans are better than central count because the voter can be given a chance to correct errors on the ballot before it's cast, but this is true ONLY if the ballot reject options in the scanners are turned on and configured appropriately. This may be done from a central point using Diebold's GEMS system with their scanners, in the same way their DREs are configured. So if someone wants to mess with a Diebold "Accuvote" system, Op Scan or DREs make little difference. They can do it because they are both configured through GEMS. DREs are still worse though.
See this post from last night about how GEMS can be used to screw up Op Scans and DREs alike:
<
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x351376#351709>So like a lot of things in life, it's not so simple.
A lot depends on the PEOPLE running the election and the PROCESS in place to guarantee accuracy. You need checks and balances to keep everyone honest. But with paperless DREs, there is not even a CHANCE of independently verifying the vote count. With any kind of Op Scan, you can do a full manual recount or random audit any time you want, (until the paper is destroyed). And before and after the election, you can use test decks of ballots to see that the machines are counting properly. Paperless DRE recounts are meaningless and you have to rely on so-called "Logic and Accuracy" tests which themselves are not verifiable.
Hope this helps.