I'm also surprised that someone, who purports to be knowledgeable on kernel development, was not cognizant of the fact that the code he was looking at was only started 8 months ago and no where near production quality. To start making assumptions on the quality of an OS based on very early code was extremely immature and makes me wonder what exactly his qualifications really are. The primary goal of the Fuchsia team is to try and get something booting up and running and then start to iterate and polish.
"I just learned by doing. Got annoyed by something being bad in one of the BSDs, found out how to get into their secret chat channel, started whining and got told to shut up and write the code. So I did."
And rambles more about why he feels Fuchsia did the wrong thing. Specifically because better implementations of some parts with licenses compatible with Fuchsia existed. Which I suspect is a nod again to BSD.
With a more calm tone I think saying that writing your own suboptimal tick timers, or locking strategy, or whatever...versus using something known good, is fair criticism. Not being my space, I wasn't sure if the base criticism was valid. The responses here seem to indicate they are.
Fuchsia is probably 2 or more years away from an Alpha release. Things are changing at a rapid pace.
If the suboptimal stuff was existing, I take your point. Perhaps it was straight from LK?
But writing something new, that has flaws, as a placeholder...when something better exists already, and has a compatible license...that seems fair game to question.
Especially if it's something that has tentacles and creates dependencies. Placeholders sometimes don't stay as such.
He seems to be reciprocating the years of criticisms and verbal abuse given to him by Torvalds.