Voting Systems with "Open Source" Software are Computerized Voting

[Wilms]

:shrug:

What am I missing? What are the advantages?

[Occulus]

1. Open source software allows for public verification of the code

Unlike the ones made by Diebold etc.

[Wilms]

3. You mean public verification of the source code.

That's not what runs on the machine, though. Object code does.

So how do I know what's been loaded on a given machine? And how do I know IT wasn't hacked? :shrug:

[slackmaster]

2. Open souce is auditable. The opposite, propriatery software, is not.

In proprietary systems the source code (human-readable language that gets compiled into binary objects), is secret.

[Wilms]

4. And the "compiled into binary objects" code is not a "human-readable language", however.

So how do I know what's on the machine? And that it wasn't hacked??

[slackmaster]

5. You can take the purported source code, compile it using the same open-source compiler

...that was used to create the binary object in question, and compare the two output files (yours and the binary in question). If they are not identical, then you can be sure something isn't right.

If you are concerned that the compiler itself contains a hack, you can check it out as well because it is open-source. On up the food chain.

[Wilms]

7. Would that work on those reported hacks where the bad code eats itself?

And even there, I'm assuming we're checking what's loaded on an individual machine. I don't think that's often done.

[slackmaster]

10. I don't believe it could be hidden if all the source for the application, compilers, and linkers

Are available for inspection.

[Bill Bored]

11. Maybe I could, but what about the little blue-haired ladies working the polls? nt

[slackmaster]

16. They are indeed the ones you need to keep an eye on

:hide:

[Wilms]

20. Agreed!

:rofl:

[Bill Bored]

12. Oh, and how do I know the compiler is the same compiler?

Do I need to compile its source code on another compiler?

[slackmaster]

14. Same way you validate the object you are primarily concerned with

Do I need to compile its source code on another compiler?

That's one way.

[Bill Bored]

18. And there ought to be a log file, and a log of what's going into the log file, and a log of that....

[diva77]

21. why try to insert complexity where it's not needed; simple collating skills

are all that's necessary

Sorting made simple!

[Wilms]

22. !

:D

[slackmaster]

23. I've never been convinced of the cost benefit of automated voting systems

:hi:

[AndyTiedye]

6. Compile the Source and Compare the Resulting Binary

do the same with the compiler and linker.

[Wilms]

8. That sounds helpful.

But there's still the ballot definition files for a given election...a fairly likely way to hack or just screw up an election. No fault of "open source" but it won't help there either.

So I'm still stuck with the idea that it's computerized voting, and that for any given machine/election I don't know what's going on.

PLUS, by being in the public view, bad guys have an easier time learning how to breach it. Correct? :shrug:

[AndyTiedye]

9. We Need to Inspect the Data Files Too

Ideally, the system uses some sort of ASCII files for ballot definitions and the like so any member of the public can verify that they are correct.

The best system would be an open-source system that generates paper ballots.
The voter can inspect the resulting ballot before submitting it.
Any computerized total would be considered preliminary, pending a hand count of the real ballots.
Hand-counting is televised and/or webcast so the whole process is open.

[Bill Bored]

13. Oh, so you're saying the post-election audit is really the way to verify?

In that case, the only benefit of open source is to un-privatize elections. That's a worthy goal, but it won't tell me who won and who lost, will it?

So I agree that there needs to be a really good audit, using hand-to-eye counts of the paper ballots, with a reliable chain of custody and ballot accounting (because paper ballots are easy to manipulate like software, no matter how they are produced). Too bad no state actually does all that, except for the occasional "spot checks" huh?

[slackmaster]

15. The easiest way to audit an election is to count the paper ballots by hand under dual control

Dual control = Two people doing the counting and keeping an eye on each other.

:hi:

[Bill Bored]

17. Agreed. As long as enough of them are counted, and they haven't been tampered with beforehand. nt

[slackmaster]

24. Controlling boxes of paper is a lot easier than controlling things nobody can see

Automated machines that count and sort currency in a bank make sense. Not so with ballots IMO.

[eridani]

19. Open source is better than private. However--

--if you want to make sure your laboratory scales are working properly, which is more useful? A schematic detailing the scale's design, or a set of standard weights that you use to check its performance?