You say this because you believe that there is a large number of people voting as their dead grandmother. However there is no proof of any kind that large numbers of people are cheating in this manner.
A solution in need of a problem.
Originally Posted by txdot-guy
I disagree wholeheartedly. First some evidence that's driving me this way:
https://chicago.cbslocal.com/2016/10...ond-the-grave/
https://strong.academic.wlu.edu/humo...ago-elections/
https://www.dnainfo.com/chicago/2016...-stolen-trump/
https://www.chicagotribune.com/opini...926-story.html
https://www.nbcchicago.com/news/loca...oters/2358435/
https://www.fox5ny.com/news/shipment...hicago-airport
Either way, this is Vulnerability Management 101. The key to securing a given system is to identify weaknesses and remediate them through fixing the weakness, having compensating controls, or replacing the weak part with a different system or process.
By and large, the two largest categories you can break vulnerabilities into, are exploiting a system to behave in a way it's not supposed to, like tampering with the voting machines, and the other is credential weakness on behalf of the users that gives people access who shouldn't have access.
Now to the first part, they argue that the machines are safe, properly set up, and properly patched. I won't get on the soap box about that, as that's one of the major contention points and all I have to work with is conjecture that it is or it isn't.
For the second part, user identification, I'd like to look at current trends in other fields:
-You want a given user to have "least access" where they can only do the job they need to do and nothing more.
-You want credentials to have a minimum amount of complexity
-You want a mandatory secondary factor of credential
-You want people involved in the credential management to make sure they're not being socially engineered, (in this case, the people who grant voting ID's and the people who check those ID's at the voting place.).
-You want to make sure that users are purged regularly so that their credentials cannot be falsely assumed.
The more modern approach to this is called
Zero Trust Architecture.
Some of these methods translate, but not all. Firstly, you have to combat the state by state differences in voting, some require ID and others do not. The first step would be making identification mandatory. This doesn't need to be a separate type of documentation, but could be rolled into Driver's Licenses and State IDs, which commonly have other provisions to them. For a second factor, biometric data, such as a fingerprint could be used, but that would require the collection of fingerprints for everyone, which not all states do.
The easiest solution by far is to regularly purge rolls and require people to sign back up. If voter registration were tied to Driver's Licenses and State ID's, one could streamline the process, making sure data for both is current.
In short, purging isn't the solution, it's the first step in securing elections.