In my precinct, when there was optical scan (prior to Diebold), I would take my ballot to an election worker. I would then stand there, watch him insert it into the optical scan machine, and he would tell me whether the machine accepted or rejected my ballot. There was a light on the scanner too. If the machine rejected it the election worker would tell me so and investigate/give me another ballot.
My understanding is that Cuyahoga county is using a 'Central Optical Scan System'. What this means is that you simply hand your ballot to an election work and at some point the ballot is transported to a central location and scanned there. This means you do not see it scanned, you are not made aware of a problem that you could correct, etc.
That is what the ACLU is suing for.
Here is a link to the ACLU Motion (pdf):
http://www.acluohio.org/issues/VotingRights/ACLUvBrunnerBriefInSupportOfMotionForPreliminaryInjunction.pdfIf you look in that motion at table 2, you will see that the 'residual vote rate' (number of votes not counted) was 1.7% when this 'Central' system was used. When a precint count method (as I described above) was used, the residual vote rate was 0.7%.
Since the more inaccurate method is ONLY going to be used in Cuyahoga county, a predominantly black district, the ACLU clams that is another unfair bias.