That is totally dependent on the use case.
I think that is @FranklyFlawless’s approach. Nothing on the phone (on the eMMC drive), so can reflash at the drop of a hat.
For me, definitely not. When I upgraded from amber
to byzantium
that had to be a reflash because amber
didn’t come with an encrypted file system out-of-the-box and I wanted to introduce LUKS (and while it is technically possible to encrypt “in place” there can be issues). So that was a very careful plan of: copying lots of stuff off the phone, reflash, and then re-establish stuff on the phone. So, for me, no, I wouldn’t reflash unless I had to.
Or let’s say that the phone is bought second-hand with therefore unknown software, settings, … on it. So it should be reflashed on Day 1 of acquisition. Or, failing that, should be reflashed sooner rather than later.
For a dissident in a high-threat environment, yep, it may be wise to avoid putting stuff on the phone at all (and hence can reflash easily at any time).
If the user irretrievably breaks the contents of the phone then, actually, it should be restored from backup, not reflashed, but, failing that (e.g. if there is no backup), yep, reflash.
So I am not going to generalise as to what all users should do.
That can be true but it can still be mucking around to re-establish. (For example, I have ?hundreds of contacts on my phone. Am I sure that I can re-import all of them? To be on the safe side, should I export them before reflash and then re-import them after reflash? Should I be using LDAP? )