ghost
commented
5 years ago You have reason, but my point of view is that the "create identity" workflow is faster if we do it like that. So this is the current workflow: 1.- ask for label 2.- ask for passphrase 3.- call relay to create id, and we get a counterfactual address and keys 4.- with the passphrase, we need to unlock the key container and then, call to the name resolver to bind the label with the eth address.
So, we can't do step 4 without ask for passphrase. Anyway, I guess that process is faster for the user like this.
Of course we could do this: 1.- ask for label 2.- call relay to create it and get eth counterfactual address and keys 3.- ask for passphrase 4.- unlock the key container with new passphrase and bind label with address
Not sure about usability... Maybe @rafal-na can help us in this discussion :)
Current: The check that verifies that the username isn't repeated happens after adding a password, which may make the user need to go back. If the user picks an existing user now, the user needs to write the password multiple times.
Expected: Having this check in the screen where you pick the username would make the interaction smoother. If the check is done before, the user only needs to write the password one time.