In talking with a number of people over the last week about the ES&S "smoothing filter" memo, some allegations have been raised by both sides which I'd like to address here. It might help clear up some unanswered questions such as -- who knew what and when did they know it?
If you've been following this story you'll know that the memo about the "smoothing filter" problem with ES&S machines languished unnoticed in cyberspace since last September. When the blogosphere picked up on it a few weeks ago, the Jennings camp seized it as "a smoking gun" and accused ES&S and Sarasota Elections Supervisor Kathy Dent of withholding the memo from them. Jennings' lawyers, as part of their lawsuit, had asked Dent's office for all correspondence pertaining to problems with the machines but had never received this memo. So why did Dent's office fail to hand over the memo and why did ES&S say in court that its machines had worked perfectly when it knew about an existing problem with the machines that had never been fixed?
Dent said that the memo was misfiled in her office and that she hadn't intended to withhold it from Jennings. She also said that it didn't matter that she hadn't given the memo to the Jennings camp because she says she had given them e-mail correspondence from ES&S discussing the smoothing filter problem with the machines, as well as e-mails between Dent and her employees discussing it. If Jennings' camp had read these e-mails, she said, they could have seen the same information that was in the missing memo. Here are two of those e-mails:
From: Cihacek, Angela [from ES&S]
Sent: Thursday, August 24, 2006 9:59 AM
To: [Numerous Florida election officials deleted here]
Cc: Bennett, Linda; Buchanan, Janet
Subject: ES&S iVotronic 12 inch screen users
After a number of inquiries from several of our iVotronic 12 inch screen users that some of your screens are exhibiting slow response times. ES&S sent out a letter on 8-15-06 concerning the issue. Attached is a voting booth instruction sign for your use. If you have any questions, please call your customer service representative, Lora Peterson, at XXX-XXX-XXXX
Election Systems & Software
11208 John Galt Blvd.
Omaha, NE 68137
XXX-XXX-XXXX Ext. XXXX
NOTICE: All mail sent to and from the office of the Supervisor of Elections is subject to the public record laws of Florida.
From: Dent, Kathy
Sent: Thursday, August 24, 2006 10:21 AM
To: IT Department; Crete, Karen; Goodell, Tom; Bain, Barbara; Fowler, Cathy; Dingess, Traci; Powell, Terrina; Walker, Bobby
Subject: FW: ES&S iVotronic 12 inch screen users
As you know, 12" screens may have slower response times. I think that has happened in at least one of our machines. I need input from you all about whether we place this poster in each booth at this stage of the game. Since poll worker training is almost over, we will not be able to go back and talk about this with them.
I do know that ES&S is trying to getting a recertification to change this before November.
Give it some thought and let me know.
When I spoke with Alec Yasinsac last week about the ES&S memo (Yasinsac was head of the team of computer scientists from Florida State University and elsewhere that examined the ES&S source code and produced a report saying the smoothing filter issue was not the cause of the undervotes) he asked me the same question -- why was the Jennings camp making such a big deal about the memo now when they knew about the information in the memo back in December? In fact, Yasinsac said that Dan Wallach, the Rice University computer scientist who is the Jennings technical expert, did have a copy of the memo in December and yet didn't bring up the issue of the smoothing filter problem during court hearings. Why accuse ES&S of not discussing the "smoothing filter" issue if Wallach didn't bring it up either? "Why don't you ask Dan Wallach that?" Yasinsac said.
So I did. Wallach says he didn't have a copy of the memo. He says he might have seen it pass by on a voting list back in August when the Florida activist first made it public. But a lot of things crossed that voting list that he glanced at quickly. And back in August the Jennings/Buchanan election hadn't occurred yet. The first time the memo came to his notice, he says, was when everyone else in the blogosphere became aware of it a couple of weeks ago.
As for the e-mails that discussed the memo, Jennings spokesman David Kochman tells me that Jennings' lawyers didn't receive the e-mails from Dent until January, long after the December hearing and only after several attempts to get them.
"There is no way we could have talked about the (smoothing filter) issue in court since we did not receive the Kathy Dent emails until January 25 -- more than a month after the court hearing and just a few weeks before the memo story broke in the press," Kochman wrote me in an e-mail. "It's also important to note that we had made several earlier public records requests and interrogatories that should have produced both the memo and the email. . . . Bottom line is, the state, the county, and ES&S all withheld the memo despite several efforts to obtain such information."
So who did have the ES&S memo besides ES&S, Kathy Dent and other election supervisors in the state who used the ES&S iVotronic machines? Computer scientist Alec Yasinsac says he and his FSU team had it when they examined the ES&S source code for the state and wrote their report.
"My experts, my folks that came here, had the letter and knew about the (smoothing filter) problem," Yasinsac told me. "We looked in the source code to try and find the way it worked. We found the way it worked, we addressed (the issue) extensively at two different places with technical findings in the report. We produced a conclusion that specifically said we believed (the smoothing filter) didn’t cause the problem (of undervotes), and that’s our answer."
The report, by the way, doesn't say that the researchers possessed the ES&S memo or that they learned about the smoothing filter problem from ES&S. It only mentions that the researchers were aware of rumors on the internet that the smoothing filter might have been the cause of the high undervote rate in the Jennings/Buchanan race.
The allegation has been floated on Internet newsgroups that the iVotronic touch screen filter could have caused the undervote. No explanation has been offered how the effect would confine itself to a single race on a single screen. The touch screen filter does not act differently on different screens. (p. 48 of the report)