You make these analogies attempting to equate an advertised open WiFi network to an unlocked home, while ignoring the precedent around both of those things.
It is expected that people connect to your advertised open WiFi network. It is not expected that people wiggle your doorknob to check if it's unlocked or not. If you put a sign on the door advertising, "the door is unlocked!" then I wouldn't be surprised when someone mistakes that for "come in".
A better analogy is:
I leave my door open with a welcome sign out the front.
Two people enter.
One of them picks the pocket of the other.
And then the thief blames the guy who told him about the open door in the first place.
While I do remember hearing about Google Maps vehicles connecting to open WiFi networks in the news, I don't recall hearing about private credentials being published. Was that the case? I thought it was just a map of open WiFi networks that was published with basic details such as SSID?
Edit: I found the article (2010, holy cow does time fly). It looks like they did collect payload data for non-encrypted traffic. Even though the data wasn't published in any way, I must agree that they went too far. I would have no issue if they were to simply verify that they could connect and record basic info such as SSID, but collecting payload data from network requests was inappropriate.
I don't find it particularly troublesome that maps of open WiFi networks exist.
I do not, however, think that it's okay to behave maliciously, or inappropriately on open WiFi networks.
My earlier response to your comment about hoovering plain text passwords didn't properly acknowledge the bad behavior that took place. I concede that you are correct, it was rude and insidious behavior.
It was never the connecting that bothered me, it was the storage of the data encountered.