Most (in my experience) "newer" Microprocessors will withstand an accidental short once or twice... to ground, even occasionally instead of once or twice. Most All Microprocessors will survive a short to Vcc once or twice but likely never occasionally and No microprocessor will survive shorting to a positive supply greater than Vcc to less than ground (Negative voltages) without a resistor big enough to limit the current to what the input protection diodes can withstand. Some types of micro's are more robust than others and for some obvious reasons the choice of the Atmel product was a wise one. The '328's are a hardy chip family. I've never yet read here that shorting an "output" pin to ground will destroy that pin or the processor instantly or even soon. I've read many times of the per pin current limits and infrequently of "other" types of shorts. I've NO DOUBT that a long term short... 5 - 10 minutes would cause enough internal heating to begin the "electromigration" mentioned and even cause the device to fail. That's one major issue with mos devices and overheating from wrong type loads is the major failure event for Mosfet power transistors (Bipolar devices fail in a similar manner from "second breakdown" which is a place on the "overloaded" transistor die that has a crystalline defect, the defect gets hot and fails causing more area's near the failure to fail until the device fails utterly. If the same transistor is never operated at or above it's rated max current these types of failures are most rare. I used to test silicon power transistors for that issue by measuring the C-E resistance... it should be a near open circuit > 5 megs. Measurements less than 50K were a danger flag for me and I replaced those parts. With a mosfet there is no easy way to look for electromigration unless it be in the gate capacitance... The defect in my experience is mainly damage to the gate structure (The oxide layer under the gate metallization) and the damage is really unnoticeable until the device fails although my high power fet use was limited to RF power amplifiers. I used many power mosfets but the ones that failed were the ones that regularly were operated at 100 + Deg C (failed fans or air filters clogged). All of the above mentioned conditions are "Over the Top" and are the exception rather than the rule.... There have been a lot of "My Pin puts out 1.5V... Why? types of questions but not many in which the poster reported failure because of the error...
Did I "Fry" my Uno?, I accidentally connected a pin to Vcc or ground or I connected my servo power to a pin on the Uno because I wanted to turn off the motor...
but rarely reports that the device utterly failed. Common sense is the key here, A short should be avoided if at all possible but isn't cause for failure or ALL of us would be seeing more messages to that effect. Great topic for "Trolling" though... I suspect someone got stood up for "Saturday Night" and is revealing his/her frustration here...
Bob