I appreciate the feedback, and we do try to adapt as best we can. There are several design decisions involved here, not 100% sure which one(s) are the critical ones here. I do think that disallowing passwords such as ‘password’, ‘secret’, ‘123’ and ‘a’ is an indisputably good design decision.
The other decisions, such as using the sign in metaphor for deciding on what password to use by default for encryption and decryption can of course be discussed, but what you can’t know is that it’s based on actual use of the “old” model. One recurring problem with the old model with the need to re-type your password every time an encryption was to be made and the only verification done was by typing it twice was that users mistyped (twice) and then lost their data.
The sign in metaphor allows us to verify that the password used is actually the same one as was used the first time. This has virtually eliminated the data loss caused in version 1 by mistyped passwords. It has also reduced the number of instances of lost data due to immediately forgotten passwords at first time use.
Finally, the design decision involved here with sharing is not requiring the sharing of actual passwords, is also in my mind a good one. However, here we’re open for allowing password sharing in addition, since we do see some cases where this would be beneficial. We have an issue for this, you can follow it here: https://bitbucket.org/axantum/axcrypt-net/issues/131/add-sharing-password .