I'm a software developer by trade. Give me a set of requirements, I'll write the software, and I'll even help you refine the set of requirements by including what should be obvious.

An election system MUST:

• Allow everyone who is allowed to vote to cast their vote and have it counted they way they intended it to be counted.
• Prevent anyone who is not allowed to vote from having their vote counted.
• Inspire confidence that the election results truly reflect the will of the people.

Because this is user-facing, and because the users aren't necessarily on the upper end of the IQ spectrum, I include the following:

• Be simple enough that someone with an 85 IQ can eventually grasp how it works to prevent fraud.

Here's my design proposal:

• We keep a database of eligible voters along with their identification. This must piggy-back on some kind of state-issued ID system, or rely entirely upon name and address.
• In order to get added to the database, you must present evidence that you are who you claim to be and that you are eligible to vote.
• We must also have a way to remove those who are not eligible to vote, or challenge and resolve those whose registration are questionable.

This first part of the voting system is essential. Note that it ALSO requires that the list of eligible voters be made public, along with their forms of ID (how they were authenticated). For instance, if John Smith is an eligible voter, then we must be able to see that he showed either a birth certificate or a state-issued ID along with some proof of residency and what that proof was.

The more secrecy we introduce in this first database the more likely it is that someone will exploit it.

The second part of the system is a little bit more complex.

• When casting your ballot, you must present identification that establishes you as a person in the voting database. We can make this a pre-requisite to receiving a ballot as well as a pre-requisite to accepting their ballot. That is, if you take the ballot home with you (absentee voting) you must again present the ID when you drop off your ballot. By preventing anyone else from handling your ballot, that prevents anyone from tampering with your ballot.
• If you aren't able to accept or deliver the ballot in person, say, because you are sick or disabled and can't leave your home, then we can arrange to have an election worker visit you and mark off your ID, etc... Of course we would record who visited you and when.
• We should have some sort of physical mark that is applied to the person's body indicating that they have voted. This prevents anyone from voting twice with separate ID. Purple ink on a finger is a great way to do it.
• We live in the digital age, so it is entirely reasonable that we capture images of the identification and such. This is not unlike companies that take a picture of the furniture after it was delivered.

Now for the counting.

• Each ballot is marked with an ID that links it to the voter who cast the vote, and thus, the identification, etc... presented. However, to ensure secrecy, the link between the voter and the vote is only known by the voter and kept hidden for any purpose except validating the integrity of the election.
• It is recorded whether or not a person voted in the voter registration database, and this is publicly available -- NOT their vote ID.
• A picture of each ballot AS IT WAS RECEIVED is captured, which contains the ID. These are publicly available. Anyone can view them.
• A record of how the vote was recorded is also captured along with the image, so people can challenge and say, "They clearly voted X when it was counted as Y!" This is also made public.
• Thus, a database of vote IDs along with votes is kept. People who know their vote ID can look it up in the database and compare it with how their vote was counted. People can also run the tallies themselves to see if they match.

This is all well and good, until people start playing shenanigans, so we need a cost-effective way to challenge election results.

• If someone shows that the numbers didn't add up right, then they can show their own additions and challenge the results. Since they are both using the same data, this is trivial to verify, and a court can handle that lawsuit in a matter of minutes.
• If someone voted who shouldn't, then you can point to their record in the voter registration, explain why they are not allowed to vote, and then indicate that they did indeed vote. The judge can look up their vote ID and then invalidate that vote. The vote DB will be updated accordingly, and the results changed to reflect it.
• If someone cast a ballot for another person, then that information will be available with the images and the person who issued or accepted the ballot. Indeed, the person who purportedly voted can say, "I didn't vote" and thus the ballot is invalidated. A judge can invalidate the vote and the vote totals updated accordingly.
• If a ballot is read wrong, then you can either get a judge to agree with you and change how the vote was counted, or get the person who cast the vote to make a sworn statement as to their intent, and get a judge to change the vote. Vote totals are updated accordingly.
• If there are ballots being counted that shouldn't be, then get a judge to agree with you and the ballot is thrown out. Totals updated accordingly.

This is a pretty fair and simple system, completely transparent. Although it cannot ensure that fraud will never occur, it can certainly make it more apparent when it does occur. How hard is it for someone to list out the ballots, voters, etc... that they find problematic and suggest their own vote totals when all the data is readily available?

And what happened in Washington State, where the vote was clearly stolen but Culp unable to fund a massive lawsuit that would require millions to challenge it, can never happen again.