AutoMARK under attack?!? I need some smart DU help with some major info

[demodonkey]

OK I need some help from the great DU research team...

"Someone" (in a position to influence several states' buying decisions about machines) has pointed out to me that the AutoMARK software is 2002 NASED certified, AND the hardware/firmware (mostly with ES & S stuff) is 2002 NASED certified, BUT taken together the whole system is only 1990 NASED certified.

WTF?... Look at the first entry in the document at link below from nased.org... Is that a misprint? Didn't I remember a few weeks ago seeing a news release that AutoMARK was completely 2002 certified for use with ES & S components for the rest of the system?

Updated List of ITA Approved Systems from 12-03 to 7-05
http://www.nased.org/ITA%20Information/NASEDQualifiedVotingSystems12.03to7.05.pdf

I need you DU researchers to dig up the press releases about AutoMARK getting its 2002 NASED (would be in the last few weeks?) Or anything else you guys can find to refute this notion that AutoMARK not fully 2002 certified...

TIA for your help!

[FogerRox]

1. here

http://www.automarkts.com/Main.htm

[Cookie wookie]

2. The AutoMark is qualified to

2002 NASED Voting System Standards (VSS). However, it is used with ES&S's optical scan, which is qualified to the lower standards, in this case 1990 VSS (all the DREs at the present time are qualified to 1990 VSS).

[Bill Bored]

3. What is "lower" about the earlier standards?

The accuracy spec went down in 2002, didn't it?

Also, the Automark doesn't count votes, it only casts them. Once the ballot is filled out by the Automark, it's scanned and counted by the same scanner as any other ballots at the polling place. So what's the big deal?

[FogerRox]

4. NO the accuracy requirements went up

votersunite has the 1990 and look at the FEC 3.2.1

In 1990 you were allowed 1 error in about 700k
NOW you are allowed 1 error in about 3.1 million.

[Bill Bored]

6. Read the actual standards:

There are 2 threads in which this is being discussed. I hope we can nail it down here:

http://www.glynn.com/eac_vvsg/vg1/v1s3.htm#v1s3t577

>>
3.2.1 Accuracy Requirements

For testing purposes, the acceptable error rate is defined using two parameters: the desired error rate to be achieved, and the maximum error rate that should be accepted by the test process.

For each processing function indicated above, the system shall achieve a target error rate of no more than one in 10,000,000 ballot positions, with a maximum acceptable error rate in the test process of one in 500,000 ballot positions.
>>

That's 1 in 500,000

Here's 1990:
http://twinpeaks.cs.ucdavis.edu/evote/standards/VSS-1990.pdf

>>
3.2.4.2.7 Recording Accuracy

DRE systems shall accurately record each vote and ballot cast. Accuracy as here defined means the
ability of the subsystem to detect every selection made by the voter, to add permissible selections
correctly to the memory components of the device, and to verify the correctness of each of these
operations. It also means the ability of the device to preserve the integrity of voting data and ballot
images (for DRE machines) stored in memory against corruption by stray electromagnetic emissions,
and internally-generated spurious electrical signals.

Recording accuracy may be achieved or enhanced by the incorporation of multiple detection and
memory elements that employ device polling techniques. Corrected data errors shall in these instances
be logged by the system.

The error rate measured by these criteria shall not exceed one part in one million, as applied
independently to the voting data memory and to the ballot image recording devices.
>>

That's 1 in 1,000,000.

1 in 1,000,000 (1990) is more accurate than 1 in 500,000 (2002), isn't it?

[FogerRox]

7. youre comparing the wrong numbers

Target rate VS minimum rate
have to compare target rate vs target rate

[Bill Bored]

10. Actually when it comes or errors, the rate is a maximum, not a minimum.

And the target appears to be just that -- a target. The way I read the 2002 spec, the machine could be certified if it does not exceed the maximum error rate and that's 1 in 500,000 or 2 x 10^-6.

I think they weakened the standard but made it look like it's tougher by stating a lower target error rate that nobody really has to meet.

[FogerRox]

14. yes max--- sorry

[Pig Farmer]

13. There's another subtle difference

1990: "votes"

2002: "positions"

In 1990, only voted candidates are counted towards the error rate.

In 2002, UNVOTED candidates are also counted.

Apples & oranges.

PF

[FogerRox]

34. details-- good point-- but if you dont have 2002 certs it may be that your

equipment cant be used in Federal elections

[Eric J in MN]

5. Article by John Gideon

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=74&Itemid=30


Does ES&S Really Want to Sell the AutoMARK Machines?
By John Gideon
May 25, 2005

The AutoMARK, developed by AutoMARK Technical Services (ATS) is a computerized ballot-marking device. The voter makes selections using a touch screen, then the system marks a ballot with those choices. That ballot is then either read by a precinct based optical scan machine or deposited into a ballot box and read later by a central count optical scan machine or by hand.

A ballot marking system designed to provide privacy and accessibility to voters who are blind, vision-impaired, or have a disability or condition that would make it difficult or impossible to mark a ballot in the usual way. In addition, it provides language assistance to voters who are more comfortable speaking an alternative language or who have reading difficulties. The AutoMark voter assist terminal has been developed with input from election authorities and disability organizations, and meets all of the requirements of "The Help America Vote Act of 2002."
ATS describes the system as:

Election directors and private citizens who are concerned about trusting electronic ballots consider the AutoMARK highly preferable to electronic voting machines, even those that provide a voter-verifiable paper record of the vote. Many prefer it because it offers a cost-effective way to comply with federal laws. As a result, many counties across the country have been eagerly awaiting the federal qualification of the AutoMARK. The system just completed testing successfully and qualification is imminent.

Election Systems and Software (ES&S) presently produces only Direct Recording Electrical (DRE) voting systems and optical scan voting systems. They have no system that completely satisfies all of the HAVA accessibility requirements and provides a voter verified paper ballot as required in some states. The AutoMARK voting system satisfies the accessibility requirements and provides a paper ballot.

[FogerRox]

8. go here Bill:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x385691#385707

[FogerRox]

9. I'm back--- had to run an errand

FEC 3.2.1 as cited in HAVA title III, section 301 is:


http://www.fec.gov/agenda/agendas2001/mtgdoc01-62/overview.htm

1. Target error rate: a maximum of one error in 10,000,000 ballot positions, and

2. Testing error rate: a maximum acceptable rate in the test process of one error in 500,000 positions.

NOw the 1990 standards are here in Appendix F.5

http://www.verifiedvotingfoundation.org/downloads/FEC_1990_Voting_System_Standards.pdf
Desired error rate = 1 in 10,000,000
Maximum acceptable = 1 in 100,000

Further more--- in F.5=
This results in the following decision criteria:
• If the system makes one error before counting 167,753 consecutive votes correctly, it will be
rejected.
• If the system reads at least 297,589 consecutive votes correctly, it will be accepted.
• If the system correctly reads more than 167,753 votes but less than 297,589 when the first
error occurs, the testing will have to be continued until another 465,342 consecutive votes
are counted without error (a total of 762,763 with one error).
This test design replaces the horizontal axis in the time-based illustrations with the total number of
trials. Just as there was a minimum time to accept without failure, there will be a minimum data sample
size to accept without error. As a practical matter, the test is terminated if an error occurs in less than
167,753 votes. The vendor is then required to improve the system.

Review:
2002
1. Target error rate: a maximum of one error in 10,000,000 ballot positions, and

2. Testing error rate: a maximum acceptable rate in the test process of one error in 500,000 positions.


1990
Desired error rate = 1 in 10,000,000
Maximum acceptable = 1 in 100,000

So the maximum acceptable rate of error is 5 fold more------in the 2002 standards

(Poke at Bill)-- DUDE you gotta read the whole thing, man----

[Bill Bored]

11. We have the same 2002 specs but the 1990 ones differ.

I need to find ot why. Mine came from a .edu site -- UC Davis.
Yours was from Verified Voting. Yours is from an appendix, mine isn't.
I have to go back and read the 1990 ones to see why there's a discrepancy.

[FogerRox]

12. No-- Both PDFs are 204 pages-- they are the same

you didnt read the whole thing thats all.
Your link has an appendix--- you just gotta read it--- look for appendix F.5

I went back and checked your link-- its there try pages 149- 150

[JohnGideon]

22. Thanks Eric...

I actually wrote this article in hopes that I could either drive a wedge between ES&S and AutoMARK or bring them together so the AutoMARK was offered more widely.

It seems to have brought both parties together and ES&S is doing a lot better job of marketing the AutoMARK system.

[Cookie wookie]

15. Just heard that the AutoMARK is

now going to be fully qualified to 2002 standards. The story is that ES&S had said they had problems with the AutoMARK and their 2002 qualified optical-scan machines, that they wouldn't read the AutoMARK ballot, so they were coupling it with 1990 VSS optical scans.

It seems that now the problems have been solved and the AutoMARK is back in testing with the 2002 qualified optical scans, which will give AutoMARK full 2002 qualification.

[WillYourVoteBCounted]

16. do you know which optical scans it is already qualified with?

Cookie Wookie, this is impt, we are hoping that some of the optical scans already in use in our state can be used with the Automark.

For example, we have alot of ES&S Optech IIIPE.

What other specifics are needed to define which systems can be used with Automark?

I thought it worked with any optical scan ballot.

[Cookie wookie]

18. Sorry WillYourVoteBCounted

I don't have the information you are seeking, but if I find out I'll post back.

[JohnGideon]

21. The AutoMARK is only qualified using....

the ES&S M-100, M-150/550, and the M-650. You need one of those os machines for the system to be federally qualified. An earlier post talked about the re-qualification of the AutoMARK so it is completely 2002 qualified. That statement is completely accurate. However, the new qualification will only be with the 2002 qualified M-100 and M-650.

If your state does not require federal qualification the AutoMARK could be used with ES&S optical scans or even Sequoia optical scans because they both have the exact same ballot style. However, AutoMARK has a marketing agreement with ES&S so it is doubtful that you could get just the AutoMARK itself so Sequoia is pretty much out.

[garybeck]

17. questions about Automark

isn't it still made by ESS? do we really trust these folks? is it just the least bad option?

my understanding is, it's basically an optical scan system with paper ballots, not very different from Diebold's. but the added feature is a touchscreen option to print the ballot which is later scanned, rather than filling in OVALs.

Am I missing something? Is there any assurance that the scanners are counting the votes correctly?

Do they have open source code on their scanners or tabulators? How/why is this better than Diebold's opscans?

Does anyone know, if the printed ballot from the touchscreen has a barcode, or if the voter verified portion of the ballot is scanned to read the votes?

thanks

[Bill Bored]

19. Trust NO ONE. I don't know about bar codes, but

the voter-verified portion (which the Automark reads back to the voter) is what's counted on the normal scanner, same as a hand marked Op Scan ballot.

Where can this go wrong? If the Automark is programmed incorrectly, it could mark the wrong ovals on the ballot. But this wouldn't be hard to test before hand.

According to
<http://www.automarkts.com/Technical%20Information.htm>:

"When a voter inserts their ballot into the AutoMARK VAT, it searches for a match to the precinct identification code found on each ballot and used by industry standard optical scanning devices."

(This is what allows you to have one Automark and scanner per polling place instead of per precinct.)

"The voter is then prompted to select the language in which they wish to vote and is able to carry out the voting process using the touch screen, a puff-sip device, or by following audio prompts along with a keypad. Additionally, there is a screen privacy option voter so that visually impaired users can be assured that their voting remains private."

Yata, yata.

"For voter verification purposes, the user may also re-insert their ballot, after printing is complete, in order to verify that their intent was captured. If not, they may simply follow jurisdiction-specific ballot spoiling procedures and restart the voting process."

[garybeck]

25. here's where it could go wrong

the scanner system is not part of the automark system. if a scanner is reading the ballots and counting the votes on proprietary software WHO KNOWS what is happening inside the machine. unless there are audits, the scanner could be programmed to do ANYTHING YOU CAN THINK OF without anyone knowing.

i'm not so much worried about the automark doing something wrong. we can give SOME responsibility to the voter to review the paper before it's scanned. but what happens inside the scanner, that's another story.

we all need to realize, without audits, an opscan system is barely better than a DRE, if at all. a vote for A could be changed to a vote for B just as easily inside a scanner as it can inside a DRE.

audits are our only friend with either system.

[Bill Bored]

28. Sure but what you're saying is not specific to Automark

which is what I thought this thread was about.

As far as the voters reviewing the paper before it's scanned, the blind ones can't do so independently without the Automark, so they will have to trust it just as they would a DRE.

[JohnGideon]

23. AutoMARK is NOT made by ES&S

AutoMARK was developed by Vogue Elections which was founded, in part, by one of the developers and principals in the bar-code developement. Vogue has now spun-off AutoMARK into a separate company.

In order for the AutoMARK voting system to work it needs an optical-scan system. The AM people decided that instead of spending the money now to develope and build their own OS they would contract with one of the other vendors who had OS. Hence, ES&S who has only a marketing contract.

Yes, optical scan machines have potential problems but those problems have nothing to do with the AutoMARK. The AM does not count votes, it does not store votes; it is only an elaborate pencil used to mark a ballot but it allows that to happen in as many as 9 languages and with full access for disabled voters.

The beauty of os voting is that you have a built in paper ballot that can be handcounted in audits and recounts. It is up to us to fight for those audits.

[garybeck]

24. OK, it's just a ballot marking device... and only half of a voting system

but isn't it being sold as a package by ESS?

http://www.essvote.com/HTML/products/automark.html


and, can't a ballot marking device by done by any PC with a printer for a lot less money? I don't really see the purpose of a custom designed ballot marking device, which is probably a lot more expensive than a cheap computer terminal.

[JohnGideon]

27. garybeck, you are absolutely right....

any county can go out and buy a PC and a printer and set-up their own ballot marking device for real cheap.

However, they need to add the ability for foreign languages (Hispanic, Vietnamese, Chinese, Tagolog) and Native American. Kaching!

They need to provide audio for blind voters and all of the controls need to be set-up so blind voters can tactily locate which button is which. Kaching!

Oh, and they need to program the PC and have all of the required technology for mobility disabilities, amputees, quadriplegics, etc. (sip-puff, joy stick, foot pedal). Kaching!

Hmmm, and voter controlled ability to change the font size, contrast control and color control for the sight impaired. Kaching!

Yes, I guess a county can do all of that but I bet it will cost them a lot more money to write the software and develope the firmware just so they can have one machine at each of their poll sites.

[garybeck]

29. seems to me, that a software company could package up a solution

for realy cheap. considering the HUGE market. if, instead of purchasing one of these Automark computers (how much do they cost?), a software company could provide all the things you mention at a fraction of the cost.

I'm somewhere between a beginner and intermediate programmer and I think I could do over half the things you mentioned, fairly easily with Visual Basic...


regardless, the fact that we still have only a ballot and no clear direction on how to count it properly, free of error and/or fraud, leaves me scratching my head....

[Carolab]

36. Audits and SECURITY in the procedures

surrounding casting, counting, recording and reporting...work with the legislatures to write procedures. The SOS does NOT have authority in this area. The legislature can write procedures into law that the SOS must then follow.

[rman]

20. NASED does not do the actual testing,

the tests are done by private corporations, and can not be publicly verified.

So the NASED certification of voting equipment is nothing but a cover for privatized control of elections. I'd assume this is not news to most in this forum.


http://www.sourcewatch.org/index.php?title=Independent_Testing_Authority_%28ITA%29

An Independent Testing Authority (ITA) is a company chosen by the National Association of State Election Directors (NASED) to test whether e-voting machines meet the national voluntary voting system standards for the United States. <1> (http://www.fec.gov/pages/faqsvss.htm)

According to the Associated Press, as at August 2004 the following companies were operating as ITAs: <2> (http://www.sfgate.com/cgi-bin/article.cgi?f=/news/archive/2004/08/22/national1322EDT0461.DTL)

* CIBER Inc., Huntsville, Alabama. <3> (http://www.ciber.com)
* Wyle Laboratories Inc., Huntsville, Alabama. <4> (http://www.wylelabs.com/)
* SysTest Labs, Denver, CO. <5>

<snip>

[Fly by night]

26. kick

[JunkYardDogg]

30. AutoMark Ballot Can be filled out (marked) by Hand

The Auto Mark is designed to be used by the Disabled and the Language needy
It has a lot of good disabled aids, which the TSx and the InkaVote plus do not have
It can be used for the Non-English, foreign language requirements
The beauty of it is,
it uses a PRE-PRINTED (in English) ballot
Each unit is $5,000
Only need 1 unit per polling site-for disabled and language needs
The rest of the voting population can HAND MARK the same pre-printed ballot
To satisfy the HAVA over/undervote issue, without using an optical scanner system on site, the RoV/County has to employ "an educational program", such as signage at the polling site
see page 52 (40) 116 STAT 1704, Title III, subtitle A.-Requirements (a)...
If the County is small enough, and wants to save money,
the ballots CAN be hand Counted
or they need not be Electronically scanned and recorded at the polling site with the Model 100 Precinct Scanner, thereby not purchasing this for Each Polling Site
The Ballots Can be taken to County HeadQuarters (HQ) and either counted by hand or
OptiScanned and tabulated
then the county would buy and use the ES&S Model 550 or 650 Central
Scanner and Unity Election Management System
or Purchase an Opti SCanner System and adapt it to read the 8x11 ballots with appropriate software.

[garybeck]

31. for $5000, there must be a better way

a $500 PC with some sofware could possibly do everything the AutoMark does.

are there any software companies out there willing to take this on?

HUGE market possibilities

[JunkYardDogg]

32. It is way-overpriced, but it has a bunch of disability Aids...

The disability features are better , so far, than anything else, i.e puff & sip, blind, hearing, motor functions,etc
many numerous things which physically normal people do not pay attention to.
There only needs to be one unit per polling site, to accommodate
the disabled and the language needs
the other 98% of the voting population can Hand Mark the same pre- printed ballot.
versus
A DRE ( $3,000 and up) at each voting station within each and every polling site.
I use the term Polling Site , instead of Precinct, because , in a large city, some polling sites cover two or more precincts.
In fighting the DRE's, it is important to recognize how much influence the disability "community" has in choosing Equipment.
Although they account for less than 1% of the voting population, they garner a tremendous sympathy influence, which can and does far outweigh many legitimate concerns for election integrity.
Also, when they appear before an elected official and/or administrative body, and pimp Diebold, the officials know nothing about the money the disabled are getting from Diebold.
Plus, when we lobby, we can't argue to disregard the disabled's needs, or we come off as detestable ogres.
So, what do we do?
We can say that the AutoMark does an infinitely better job of serving the disabled and language , so that Diebold is not purchased.
Then, we have to fight the ES&S tabulators. Which is again, a problem.
Now, we are waiting to see the Accupoll.

[FogerRox]

33. whats his name-- Dickson-- ? took 20 k from diebold

[Carolab]

35. Lots of stuff about mark sense machines

http://www.cs.uiowa.edu/~jones/voting/optical/

But does not specifically cover AutoMark