Was TRIAD testing out new Total Election Management System?

[many a good man]

Photo of headquarters of Triad Government Systems. From the Xenia Economic Growth Corporation :



BLURB

In addition to the ballot tabulation program, TRIAD GSI offers a Total Election Management System, which includes a complete Voter Information System. As TRIAD DSI propels into the future, they continue to strive to expand the capabilities of their software programs using the wealth of knowledge that comes from experience, and support of customers with a determination for excellence.

*****

INFERENCE

Cheryl A Bellucci, allegedly the daughter of TRIAD's founder, posted some questions between January and September this year on a forum for computer programmers that may give insight to the project she was developing for this year's election. TRIAD has been in the tabulation business for 25 years but only since 2000 has it expanded into "Total Election Management." If we can assume that Ms. Belluci's questions to the Visual FoxPro (VFP) programmers' forum, www.foxite.com, are related to TRIAD's new election software, we can conjecture about how the program is meant to function.

We can infer that the objective of the FoxPro program is to connect to remote machines and process in parallel large amounts of data. Her goal was to get VFP to run multiple processes so it can both read from and write to multiple databases at the same time. It is entirely conceivable that it could connect to these databases (on the central tabulator at each polling place) through a modem bank.

QUESTIONS

I wonder if TRIAD was contracted with the BOE for tabulation only, or if it was for the more expanded "total election management" service. I would really like to see the results of the security audit it must have underwent. Was the code frozen following certification? What methods of validation and auditing were agreed to?


BACKGROUND

TRIAD is the company whose technician illegally tampered with a tabulator in front of witnesses. The company claimed it had to modify the program specially for the recount so it would only count the votes for the presidential race, instead of tallying all races on the ballot.

Folks, computers can add numbers really fast; there's no reason in the world they would have realistically needed to modify their program...The tabulator was powered down and I bet you the tech had to replace the modem so it could receive instructions to "un-rig" itself. By Ohio law only 3% of the ballots at each pre-appointed precinct had to be hand counted. If that small stack of 3% of the ballots matched the tabulator's total, no further recount is mandatory.

There's something fishy in the county of Delaware...

[msongs]

1. ok so where's the proof she was posting in these forums???

(edit to say posting in the alleged computer forums)

inquiring minds want to know.....

Msongs
www.msongs.com/political-shirts.htm

[Carolab]

3. Here

http://www.foxite.com/archives/search.aspx?q=triad

[NJCher]

2. who thought to check out her posts?

Clever.


Cher

[many a good man]

4. S/he's offline now

S/he PM'd me in a different thread. I'll acknowledge her when s/he gets back to me. The questions posted in my "questions" link above are from Cheryl Belluci, who I am told is the owner's daughter.

Someone should investigate how this guy got the contract in the first place. Does that windowless software sweat shop look it should be the place that decides the president of the United States?

[roseBudd]

12. I found Cheryl's posts when I googled Triad, some other DUer found

an obit for the mother-in-law of Triad founder Tod Rapp, so we know that Cheryl is the married daugher of Tod, sister of Brett & Dwayne. Cheryl is married to a taxidermist named John Bellucci, who used to post and get into flame wars on a taxidermy forum. Apparently taxidermy is not real lucrative as another taxidermist speculated that taxidermy wives must have good paying jobs. Cheryl and John also own horses. Cheryl is also a Quantum Leap fan.

Dad Tod wrote the original Triad tabulation program. Perhaps in FoxPro?

Triad has both tabulation and VIS products, here is a map of Ohio counties and what kind of business they do with Triad.

[many a good man]

13. Triad's ElecTab v1.s not written in Fox Pro

We have information on other threads that claim that TRIAD's tabulation software is written in a legacy programming language called COBOL which runs in MS-DOS 6-compatible operating systems. Therefore it can run on 15 year old PCs that can't even run Windows.

This software reads the results from a standard card reader, tabulates it and presents the results in a report.

Therefore her questions in the forum are not directly related to the tabulation software. It probably concerns another project she is working on. It could conceivably be the front end for a program run in the TRIAD offices that can go out to every machine to update the ElecTab software or collect the results. Of course it can only update a tabulator PC if its connected to the internet or the phone system via a modem.

The software she is writing in Fox Pro could also be for one of the companies other products. We know TRIAD also makes the software some counties use to manage their voter registration rolls. They also have some vaguely defined suite of products they call "Total Election Management System."

In my professional opinion Fox Pro is an idiotic choice for any serious kind of software development. The only plausible reason to continue developing software with it is because the program is already written in it and its too much work to convert it all to a better language. Definitely a very poor choice if you want to hack elections. OTOH, its such an appallingly bad choice no one would expect it to be used to perpetrate fraud...

[roseBudd]

14. Father Tod wrote the original tabulation product

its such an appallingly bad choice no one would expect it to be used to perpetrate fraud...

Sometimes the least obvious is the most obvious for that reason...

Actually another DUer found an Indiana SOS site that listed Triad and FoxPro. I'll dig it up.

[roseBudd]

15. Triad and FoxPro mentioned on Indiana SoS site

http://www.in.gov/sos/elections/hava/928,2004Newsletter.pdf

Triad Foxpro 7 Boone Blackford Howard Lawrence Monroe
Putnam Tippecanoe

This PDF has to do with converting to a common jury management tool. Triad FoxPro is one of 27 conversion groups in the state.

Note that signatures will be imported as images.

http://co.howard.in.us/comm/2001/comm_010319.html

IN THE MATTER OF A COMPUTER SOFTWEAR REQUEST FROM VOTERS REGISTRATION:



Ann Spicer and Carol Shallenberger, from Voter’s Registration, were in attendance to present a request for the Automatic Precincting Module Software available from Triad Governmental Systems, Inc. This software enables the system to:

· Automatically assign precinct numbers to voter record

· Have the ability to update street segments

· Have redistricting capabilities

· Print Street segments by precinct or county

[ProudDad]

5. I checked her posts

Fairly benign stuff on the face of it but...

Sounds like she's using Visual Foxpro 6 program(s) to read/update data in Access tables.

Unfortunately, it's likely that the actual voting machine programs are written in some M$ language (not Visual Foxpro) that writes to Access databases. Bad move but typical...the Access data engine is crap and the databases are prone to damage and easy to manipulate with external programs. I wouldn't program a voting machine that way.

The real solution to the voting machine problem is to have the same machines in every precinct in the country, all software Open Source so it can be examined by the community and with a complete, re-viewable paper trail. Also, the people who setup and supervise elections should NOT be partisans -- being a campaign manager for any candidate should be grounds for immediate recusal.

Even better, vote by mail like in Oregon. No need to "guess" how many machines each precinct needs as they "did in Ohio".


:nuke: bush

[DrDebug]

9. The irony is that Foxpro is as close to Open Source as you'll get

I used to work with Foxpro programs and had a disassembler back then and it was really cute, the whole source code was still available from the executable including the variable labels and much more. The only thing which didn't get stored were the comments.

Then again disassembling the executable is crime, but everybody with a voting machine can in theory browse through the sourcecode LOL

[nickdw]

6. go here for Foxite posts backup

Here's a zipfile and tar file each containing all 102 messages (101 of which have to do with Cheryl Bellucci) in case some freeps at Foxite decide to 404/disappear these pages.

http://www.rootshell.be/~ndwinl/triad_CB-VFP.zip

http://www.rootshell.be/~ndwinl/triad_CB-VFP.tar.gz

Merry Christmas!

[megalith]

7. Hard Info on Triad - An Xmas present

Hello
I am the SE Ohio regional Co-ord for the Green Recount, Hocking is my county and I have spent hours interviewing Ms Eaton. I was also present at the BoE meeting this Monday at Hocking. This is what I know about the Triad program <ElecTab V 1.s > that runs the card reader. And note: the card reader can be any manufactuers that adheres to the original IBM standard from the 40's-50's for punch cards. Triad builds the data aquisition/reporting software that runs the card reader, not the card reader.

Electab V1.s is old code, written in Cobol and runs under DOS, it is not Windows code. It is primitive. The machine running it varies from county to county, in Hocking it was a 14 year old Dell 386 .5 meg ram, running DOS 6. In Hocking the computer running the software is not connected to the BoE local ethernet. It does not have a modem. File transfer is by floppy disk. Printed reports on vote totals are produced from the printer attached to the tabulating computer. I do believe that other installations of Electab V1.s MAY exist on other machines in the office, but cannot confirm. Triad also provides the voter registration database for Hocking as a separate product. This is the data base that may run under more modern operating systems and applications <FoxPro as an example> I do not know anything about Triads Total Election Management Software.

I have been told by another county BoE <not Hocking> that every BoE has a T-1 line installed from the state. In my sources county the BoE is tight about security, that T-1 IS NOT connected to the local BoE's ethernet. And their tabulator is not connected to the local eithernet. However this particular BoE's tabulator does have a modem. To use it they have to unplug their fax line and plug the modem in from their office. This is occationally done to upgrade the Electab software. We are told by Triad that the same version of ElecTab is in use in all Ohio Triad counties. Triad also allowed us to print the directory tree for the Hocking computer.
Triad does take floppy copies of the tabulator count files on the night of the election or shortly thereafter. Remember that this data is what the BoEs hand out the night of the election, so it is public record. However the application required to read the data <Electab> is not public. Handy data to have.
Triad maintains open web space on their server that local BoEs may use to post local election results. Triad claims that they do not post the data, but that the local BoEs do the posting.

What does this all mean? That installations and practical security vary from BoE to BoE.


About reprogramming the tabulator.
Ohio has for many decades insisted that in recounts, only the race to be recounted will be tabulated. They do not want any found or corrected votes in the recount race to affect down ballot races and throw them into question. <A tacit admission that there are errors in vote counting :) > Again this is state, not Triad policy. Triad contracts with individual BoEs, not the state. As part of their standard sevice agreement, they create at no separate charge a new SETUP FILE so that only the race being recounting is tabulated and reported. I believe that this is policy for ESS optical scans as well.

About the recount:
You take a precinct(s) that combined, total at least 3% of the county's vote total, handcount that precinct(s), then run that stack through the tabulator. If totals votes match exactly then the rest of the county is run through the tabulator and machine recounted.
If totals on the 3% sample do not match, you handcount the entire county. Which is what every one, except us, wants to avoid.
I think only one county has gone to handcount, and they invited it, because they found a major accounting error in their official total.

About "Un-rigging" via modem.
Actually any hack that affects the official tabulation of votes must also be present for the recount, because the whole point is for the recount machine tabulated totals to closely match the official totals. So the hack code <if any> must be resident and active for both runs. Additionally it must anticipate which precincts are likely to be preselected for a 3% handcount, and NOT hack the totals in those precincts. Additionally it must pass the test decks without applying the hack, as these are handcounted also. A tall order but possible. The point being that once resident and applying the hack, the script must stay in place and remain active. There is no reason to remove it. Got it? :)

About illegal "tampering" by the Triad tech
Fact is, batteries fail and in old systems they will take down the BIOS with them. I have worked on hardware since before the first IBM PC and have had this very thing happen. When it does the system no longer knows the track and sector assignments for the HD and thus cannot read the HD. Fix is to open the case, read the HD serial number, find someone that has a database for that series of drive and re-setup system BIOS with new track and sector assignments, the data on the drive is not affected. And that is in fact what happened and what the tech fixed. Happened again on Monday during the BoE meeting in Hocking. Old, beat equipment.

About the "cheat sheet"
This is far more intersting, as it indicates that there is a tacit assumtion amoungst all parties that:
* BoEs will preselect their 3% precincts <in plain violation of Ohio statute which requires that the selection be random, for obvious reasons>
* That the precincts selected will be those most likely to scan well by the tabulator.
* That the tech was trying to point out a way that BoE ballot handcounters could remember how many over and under votes are in the precinct to be handcounted as part of the 3%, and thus anticipate and make sure to find them. This is important because it is undervotes that punchcards are most likely to miss. If missed the county goes to a hand count.

Given these assumptions, the clever evil green man intent upon hacking the system will stay away from precints likely to be selected for the 3% count, and instead focus on those precincts least likely to be hand counted to apply the hack. Focus on these and you have a way to hack the results, provided you can predict which precincts will never be pre-selected for the 3% handcount.

Remember, this does not mean that the system is hacked. But it provides a rational for how it could be. Got It? :)

Lastly, there is more to the logic that any hack script MUST incorporate, given the presently observered behavior of the system. Knowing this it should be possible to construct a logical test run protocol that would defeat any logically possible hack and prove that either:
* the count is honest OR
* The count is hacked

That's what I am working on. Preparing to make history or dash a lot of hopes, but seeking the truth of the matter in either case.

Best and Brightest Soltice to all.
Sure in the promise that the light will return in these darkest of days.

Orren Whiddon
SE Ohio Co-ord
Green Recount

[merh]

8. Thanks for the post, I am starting a thread with this for all

to see & discuss.

Go to this thread for questions and discussions of the repost.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x195375

Merry Christmas to you and thanks for all of your hard work!
:yourock:

[mirrera]

10. Wrong! (reply to original poster)

Just to clear something up... The Law is if 3% of the hand count matches the machine, you can use the machine to re-count the REST of the county. If the machine and the 3% hand count do not match, you must hand count the rest of the county. Either way the whole county gets re-counted.

I started another thread a couple of days ago wondering about whether or not they may have installed a sort of second stage vote rigging program that would run clean for the 3% and then go back to rigging as usual. I.E. 1 for you plus 1 for me equals 2 for me.

The reason I came have this thought is because the WHOLE county was going to be done. So if they just un-rigged the machine, the 3% test of the machine would match, but the count for rest of the county would also come out honest.

My next thought was hmmm how would they know how many votes to run clean? Different populations in different counties would mean the 3% would be varied.

That is easy, the B.O.E. said they weren't going to hand count any sample over 500 votes regardless of what 3% came out to be. That would make it easy for a program to be patched so it ran clean for 500-600 votes, allowing the 3% hand count to come out right, and assuring them of a machine count for the rest of the county. They wouldn't care if Kerry gained just a few votes.

That said, there is no proof of any of this yet! They may have not wanted to do the hand counting for the rest of the county because it is Christmas and they are sick of the election.

All we have is motive—not of the average election worker who knows nothing— but of Blackwell, and opportunity. They can't prove they didn't rig it at this point.

Let us also not absolve ourselves of this whole fiasco. We, US citizens, allowed computer voting to get to the point where election officials depend on these vote machine companies to carry on an election.

Anyone who works in an office that uses windows machines, usually has a tech crew. These election workers are probably relieved to not have to deal with the computerized aspects of the vote. They do not see any of this—technicians coming in and "helping"—as wrong or illegal. That is our fault for letting the system become entrenched in what is most likely illegal behavior.


Edit: I am such a slow typest that by the time I finished this the poster from the re-count, before me, appeared with a GREAT explanation!!

[many a good man]

11. Un-rigging not necessary

See my post #18 in this thread: http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x195375#195483

If you knew ahead of time which precincts would be hand counted, you can program the software to count those precincts accurately and fudge the rest. That way the initial machine count and subsequent hand re-counts for those designated precincts will always match exactly.

The first and subsequent machine recounts for the entire county will also always match, as long as the algorithm for hacking the non-hand counted precincts is not changed.