It’s not necessary in that the inverters need to support it (they could just do randomly staged power ons, and the first one could pick the phase).
However, there’s a problem in addition to electrocuting utility workers (the ones around here assume islanding during outages, so that’s less of an issue now).
Say you plug the microinverter into a 16A 120V outlet in the US, and the power goes out while you’re running a 240V 40A clothes dryer. The island is definitely going to collapse at that point, and might do bad stuff on the way down.
Even in that weird non-existent corner case instance, it is a necessary function because a hypothetical microinverter is completely and utterly incapable of kickstarting a grid.
You could totally have half a dozen microinverters scattered around your household, connected to each other but temporarily not to the neighborhood distribution network.
A hypothetical microinverter is completely capable of kickstarting the "grid" within a single house. You can try it yourself: get a 2000-watt car inverter, connect it to a car battery, turn off your house's main circuit breaker, and plug a suicide cord (male to male) between the car inverter and one of your house's outlets. The electrical appliances in your house will work fine unless one of them starts trying to draw more than 2000 watts (or, as hedora says, unless you're powering only one phase of a split-phase system). This is dangerous: if someone turns the main circuit breaker back on while in this setup, the car inverter will probably be destroyed and may start a fire. But it's far from a technological impossibility. It's achievable even with redneck duct-tape-and-baling wire engineering.
"A hypothetical microinverter" could contain plastic explosives rigged to blow up your house. Hypotheticals are fine but to be clear you've drifted away from the topic - balcony solar panels with built-in microinverters, purpose built with arc and grid detection mechanisms. But yes do not rig a car inverter to your home outlets.
The grid-tied microinverter cannot do this on its own. It is incapable.
A person can add other things to the scenario and the microinverter may elect to play ball, but it is simply incapable of kickstarting bloody anything on its own.
Are there any other wildly imaginative corner cases you'd like to explore that have nothing at all to do with balcony solar, as it is implemented in Germany?
I was rebutting your incorrect assertion in https://news.ycombinator.com/item?id=45487471 about what hypothetical microinverters were or were not "utterly incapable" of. The comment you have just posted is irrelevant to that.
The design features of currently existing microinverters are irrelevant to hedora's point about what microinverters could be designed to do, except insofar as they constitute an existence proof. (Obviously something is possible if it's already being done, but the inverse is not true.)
However, there’s a problem in addition to electrocuting utility workers (the ones around here assume islanding during outages, so that’s less of an issue now).
Say you plug the microinverter into a 16A 120V outlet in the US, and the power goes out while you’re running a 240V 40A clothes dryer. The island is definitely going to collapse at that point, and might do bad stuff on the way down.