> Do you really think a homebrewed self-update mechanism is superior to the battle tested F-Droid? I don't think it makes a practical difference. > both users

> Do you really think a homebrewed self-update mechanism is superior to the battle tested F-Droid? I don't think it makes a practical difference. > both users have to be online at the same time for their messages to be delivered to each other Not only online but one has to be directly reachable, e.g. ping $IP works. With mobile connections it's rarely the case. > on my own hardware where I am in control of the logs That still means the users of your server trust you, you've just moved the problem. It only solves the problem for you as a user. > This distributes the security risks across hundreds of operators in many countries with various data extradition laws. I don't understand this argument: if a piece of (meta)data goes through one server and you think it's bad because this server can monitor this piece of data, then having multiple servers with various levels of accountability is arguably worse. > those are all really convenient excuses for an argument which allows him to design systems which serve his own interests You are still not discussing why his reasons are bad according to you, so it's hard for people that have found the blog post convincing to change their mind. I'm also curious as to which interests you're referring to, especially when we're talking about a non-profit that develop FOSS software.