Thank you for the compliment. I try to write as well as I can.
I did talk to the woman later to get my man purse back. She straight up ruffied me. I have no idea for what purpose.
As for Uber seeing stuff, I have some posts below on the topic. I think this is not a difficult problem to solve.
Actually, what it really boils down to is CAP theorem. Uber is just making the wrong choices about CAP.
In this case we aren't talking about a networked system in the normal way we do with distributed database systems. We're talking about a network of people. How do you handle it when there's a partition event? Uber's default choice is fuck it. Write one part of the transaction to the database and charge the user a disconnect fee.
That's not in anyway the correct way to handle a network partition. That's like if you have an application that receives a message and logs it, but can't get a response to my callback, and it just says, "Okay. My work here is done." And then makes a new entry in a table that says the failed callback owes 5 bucks.
That's absolutely stupid and wrong. Network partitions are going to happen. All the damn time. And if your best mechanism for dealing with them is blame the callback . . . you have some serious thinking to do. Because that shit is broken.