Handcounting is more accurate than opscan only if--

[eridani]

--you are counting a single race. Opscan overtakes it when you are counting three or more. As Land Shark is always saying, it's about the checks and balances, not any particular method.

That means that Democrats have to watch Republicans and vice versa with hand counts. And it means that we need serious and truly random auditing of opscan counts.

We have yet to have a demonstration of a central tabulator opscan hack during a real election, though it is certainly possible that this has happened. In WA in 2004, we had a rock-solid demonstration that central tabulators were not hacked, at least for the governor's race. Not only were two machine counts and one hand count within 0.01% of each other, but the hand count added votes to the totals of all three candidates, which is the expected direction of the deviation of handcounting from opscan. (Too bad that the difference between the candidates was also 0.01%, meaning that it was not possible to tell who really won. We just went with the legally required best method for single races, namely the handcount result. And also too bad that the touchscreen results in two counties could not be recounted.)

I would argue that opscan (with open source software, of course) plus extensive handcount auditing is our best hope for clean elections, simply because using two different methods in and of itself produces checks and balances. It's like counting the rings of a fossilized tree by hand and also doing carbon-14 dating with machinery. Paleontologists are very happy campers when those two results agree with each other.

[wtmusic]

1. Wha...a non-absolutist?

Right on eridani. :thumbsup:

[BeFree]

2. Is that in 550?

That opscan has to have open source coding? Where is that written?

I could be in favor of such a system. Is there any such system on the market?

[MH1]

33. Reads that way to me.

Section 2 (c):

(8) PROHIBITION OF USE OF UNDISCLOSED SOFTWARE IN VOTING SYSTEMS- No voting system shall at any time contain or use any undisclosed software. Any voting system containing or using software shall disclose the source code, object code, and executable representation of that software to the Commission, and the Commission shall make that source code, object code, and executable representation available for inspection upon request to any person.

http://thomas.loc.gov/cgi-bin/query/z?c109:H.R.550:

[pauldp]

3. What does truly random auditing mean?

Just curious, what is the best way to select precincts at random?
The lottery air-ball machines or close your eyes and point or what?
Also what percentage of total ballots counted by hand would constitute
"extensive hand count auditing"? 5%, 10%, 25% ??
I tend to agree with your stance as far as Opscan goes, and I don't spend
a huge amount of time in this forum but I've not seen specific positions
on these things and they seem rather important.

:shrug:

[kster]

4. Hand counting is more accurate than optiscan, IF

we put these secret vote counting machines on the back burner, and we start counting our votes by hand until these corrupt Sons of B*tch*s in the electronic vote counting business PROVE that they can be trusted counting our votes.

[Kelvin Mace]

7. 100 million votes cannot be counted by hand accurately

Sorry, it just can't.

Bringing in tens of thousands of people to hand count ballots dramatically increases the chances of errors, lost ballots, damaged ballots and cheating.

HR-550 requires paper in ALL elections. It doesn't prohibit hand counts, which is fine for Mount Airy, NC, but will NOT work for NYC.

[kster]

9. So we have to rely on corporate vote counting machines

because we are dumbasses. I have a little more faith in the American people than you do, WE CAN COUNT 100 MILLION VOTES, WATCH US.

[eridani]

11. You'd prefer relying on Republicans handcounting behind closed doors?

Not me. Checks and balances.

Public fire departments buy their fire trucks from private companies, no? The reason that isn't a problem is that the specifications and operation manuals are all open book--no secret information. Make that the case for opscans, and the ownership problem disappears.

[BeFree]

15. Is 550 going to get that open book published?

[eridani]

24. It's a start

I'd call it a floor rather than a ceiling myself.

[kster]

17. I'd prefer Democrats and Republicans counting the

ballots and recording the totals in front of plenty of observers at the precinct level.

[diva77]

20. Here, here, kster!

:toast:

Precincts are discreet numbers of no more than 1200 voters (I think nationally); so creating an infrastructure that contains (hand)counting at the precinct level and thus decentralizes the count, will protect us from centralized control over the results by 1 RoV or some SoS (or higher gov't official) who has the ability to override results on, for example, a GEMS tabulator from a remote location.


Everyone makes handcounters sound like a bunch of bumbling fools. Handcounting is a skill that can be improved upon with practice! Besides, studies are accumulating which show that hcpb yield results that are MORE accurate than machine tabulated (will post when have actual info. to cite; shortly)

[Kelvin Mace]

32. Have you ever served in a precinct?

Actually WORKED an election? If not, try it a few times, then get back to us.

Go to your local board of elections. Tell them you want hadcounts only. When they stop laughing, tell us what they say.

[eridani]

22. But just saying you want hand counting doesn't get that done

What about precincts in which Repubs heavily outnumber Dems or vice-versa? No one has ever called Warren County, Ohio to account on its fake "terrorist" alert. How would Warren County's results be improved by handcounting under those circumstances? More different methods add checks and balances. Opscan gives the hand count something with which it can be compared. I'd suggest adding a third--weighing ballots as proposed by John Washburne. I've actually checked out 80# paper variability, and weighing would be quite doable.

[Kelvin Mace]

31. There you go again,

injecting reality into the argument to people who refuse to deal with the real world.

[Kelvin Mace]

26. No, you can't

and it has nothing to do with being dumbasses, it is simple reality.

You will will need tens of thousands of people nationwide to count that many ballots in a timely fashion. Who will vette all of these people?

When you have this many people handling the ballots, the security and accuracy goes DOWN, not up. The more people handling the ballots, the more chances for error, damage, lost ballots and cheating.

What is so hard to understand about this?

I swear it is like talking to intelligent design advocates.

[Kip Humphrey]

25. Boy, someone should tell the Germans about this...

[Kelvin Mace]

27. Try and pay attention

The Germans don't have complex ballots like we do, with 30-50 ballot items in some elections. Also, they have a standard ballot which is the same nationwide, we DO NOT.

[Kip Humphrey]

36. Hmmm... Sounds like a reason for Federal legislation...

[eridani]

13. What we will always have to use instead of proof--

--is CHECKS AND BALANCES!!! There is no such thing as aome kind of "proof" that will exempt us from that requirement.

[BeFree]

16. That would mean...

...audits are the checks and balances?

Where might there be a clear and unambiguous detailing of audit procedure that the experts agree will more than suffice?

[eridani]

23. I'd leav that one to a panel of impartial experts n/t

[BeFree]

34. Your reply...

...is clearly ambiguous. I need better. We all need better than empty platitudes about how great machine counted votes are.

The history of the machines thus far demands a full and complete examination leading to clear and unambiguous details of how the problems will be corrected.

Don't you think?

[eridani]

40. It isn't about machines. It is about checks and balances

Opscans have clearly decipherable paper records which can be handcounted for audits. This amounts to having two different counting methods, which is in and of itself a kind of check and balance. Add sorting and weighing ballota for yet another method.

[elocs]

5. I like the opscan ballot we use where I live in Wisconsin.

It is very simple--just connect the ends of the black bars by the candidate's name. In fact, the absentee ballot used here is simply an opscan ballot that you mail in or turn in. These are easy because you would almost have to be trying to screw them up and they also present an easy to read hard copy to hand count if necessary.

[Kelvin Mace]

8. Exactly!

But these people keep on dreaming that you can hand count an entire federal election.

Not.going.to.happen.

[Kelvin Mace]

6. Please don't confuse "open source" software

with the HR-550 requirement that code be disclosed. "Open sourcs" has the specific meaning of a community developed software package whose source code is openly available and freely licensed. HR-550 does NOT require OS software, it requires that propriatary software be disclosed to election officials for security and accuracy purposes.

[BeFree]

10. That's right

But then there are many loopholes that will be made open in that procedure of disclosing the code. In fact, there may be as many as 10,000 different codes needing to be examined.

Tell you what, get rid of DREs completely, leaving only opscan that has an open code with no memory cartridges, make them precinct based with an additional paper trail detailing the additions of each ballot (so that we know the machine is reading ballots correctly), and have all this written up in HR 550, and you got yourself a deal^#.


# We may think of more later items later, but that is a minimum.

[eridani]

12. You surely aren't expecting that there will come a day--

--when we will have laws that will by themselves enable us to at last trust any system, are you?

CONSTANT VIGILANCE!!!
--Alastair "Mad Eye" Moody

[BeFree]

14. No, but we can have the best laws

Then we can evaluate systems that meet that law and come to a consensus about which system is the most trustworthy.

But to me, at present, HCPB is the most trustworthy system. HCPB is the system that is used to correct machine errors, eh?

[rumpel]

18. Can I throw in my newest concern into this?

Just because I have not quite heard of this before, perhaps

But they used an electronic pollboook to verify voter registration and signature.

Any comment on this? Just curious.

[Kelvin Mace]

30. That's a whole other argument

separate but equal from the current one.

Poll books become a neat way to purge voters. Set you question up in its own thread.

[Kelvin Mace]

28. Not.going.to.happen.

But you are certainly welcome to get a bill introduced to just that.

Don't like 550? The get a congressman to introduce a bill with all the goodies you like.

Please, you have lots of allies. Write the law up and show it to the rest of us.

[MH1]

35. Ah.

I think I made that mistake, sort of. To me "open source" has multiple meanings and in this case I interpreted it as the generic meaning, that the source code is open and available for inspection. H.R. 550 actually does require that:

"the Commission shall make that source code, object code, and executable representation available for inspection upon request to any person."

My day job is actually in information technology (sadly, not in open source technology); I generally assume that when most laypeople say "open source" they mean that anyone can see the entire code, so that experts will be able to voluntarily review it at will to identify vulnerabilities or evil code.

It's probably important to recognize the distinction though.

[BeFree]

37. Yes

The distinction is that the 550 allowed code will still be private where open code means that it can be taken and altered at will with no copyright prohibiting such actions.

The problems with code are numerous. There can be an egg therein. What's an egg? A snippet of code that can hide itself until a certain date, crack its shell, come out and do it's dirty work, then evaporate without a trace.

There is nothing in 550 that even pretends to ameliorate such eggtivity.

Then, how would one go about examining all the different (10,000 +) codes? Problems, problems, problems.

I haven't seen any solutions to these problems, all I've seen is the real computer security experts telling us that computers should not be used in elections.

[MH1]

38. Huh?

If all lines of code are accessible, how can some "snippet of code" "hide itself"?

I thought I understood how "easter eggs" work in code, but I confess I am not aware of how one can "hide itself". Of course it can self-destruct (delete itself) on a certain date, but how can it appear from nowhere in the first place? There must be some code somewhere that starts the process. If all of the code is available for inspection, then it can be found. Or am I wrong?

Seriously, can you point me to a reputable technical site that explains how this is done?

Edit to add: "No voting system shall at any time contain or use any undisclosed software." This is what 550 says - see my post upthread. Regulatory agencies (FEC?) must implement a regimen that enforces that, or they have failed to comply with the law.

[BeFree]

39. I'm, no expert

But reading about all this on DU for three years now, and following hundreds of links, has brought me to a point to question what is being discussed here. Why don't we wait for that answer sure to come from the expert wandering these threads?

Lacking that, search for Avi Rubin's, Rebbeca Mercuri's websites. I believe you will find out about eggs from them.

You wrote:
No voting system shall at any time contain or use any undisclosed software." This is what 550 says - see my post upthread. Regulatory agencies (FEC?) must implement a regimen that enforces that, or they have failed to comply with the law.

There you have it. The HAVA law has been bastardized and the agencies have failed to enforce that law in numerous ways, so continuing to rely on them to do the right thing will prove to be the wrong thing, and 550, I fear, will be just another law that will not be enforced.

We need to rip the power over elections from the private corps and the agencies, and take elections into our own hands. The people's hands.

[sandnsea]

19. People finally coming around?

Optiscan counted in one location at the county level, one transmission to the state election office. 3% ballot audit. Source code in escrow or open for inspection. Numbered registration receipts. Vote by mail or ten day elections.

Been saying this for two years. But I was a freeper traitor for not supporting a hand count only system. By some people in this very thread.

[Awsi Dooger]

21. Don't forget error protection with the optical scan

It loses most of the benefit unless over votes are precluded, and under votes prompt a warning, to make sure that's intentional.

BTW, I'm confident you can regain freeper traitor status fairly quickly around here.

[Kelvin Mace]

29. And you can still be called a traitor

by the Intelligent Hand Count crowd. :)

Join the club

[eridani]

41. Bottom line on 550--

--it should be regarded as a FLOOR, not a CEILING. It would be a great idea to continue to work for stricter standards at the state level.

[truckin]

42. Excellent points on Opscan and HR550. I totally agree that

Opscan, with the proper controls, is the best voting system. You are also right in that HR550 is only a good starting point and there will still be a lot of work left to do if it passes.