Yesterday I put together a community meeting for the Persona project, which is an authentication system for the web that allows site owners to "outsource the handling of passwords" and implement a highly usable log in system in a couple hours.
The topic was "What I hate about Persona", and my goal was to subvert our over-loaded issue tracker and flooded email inboxes, to attempt to name some of the top reasons that cause people to not use Persona everywhere.
The full notes for our conversation are available in an etherpad, and this post attempts to summarize each distinct concern so that we can see these key problems more clearly, and end them one by one.
Announcing Persona Office Hours
Before getting into the list, one conclusion from this meeting. It was awesome. It's one thing to get an email from a interested community member, it's another to hear her voice as she describes the reasons why she will not be using Persona for her current project.
To keep this awesome going, Dan Callahan is going to organize monthly community office hours that are laser focused at users of the Persona platform. It's up to him to define the structure, but early conclusions are we're going to do this monthly and experiment with different tools and formats to make it trivial to participate, give feedback, and get live help from an actual human.
Naming Persona Pain Points
Here are the issues that were raised, in no particular order. I will run a survey to try to understand relative signifigance of these problems.
Also, please note that many of these issues are not new to the Persona team. We are actively working on many of these issues. This call, and this post, is an attempt to push ourselves to challenge our priorities, and be more accountable, and get these issues addressed faster.
- The persona Popup is not always working. It's just showing up blank.
- "Persona" the name is confusing. Googling it is not helpful.
- Pseudonymous identities as a core product feature. This is the notion that users typically have a "junk email" address, and we should build directly into Persona mechanisms. Also, If a user simply creates one email per site, the UI becomes unwieldy.
- The
watch()
API is hard to understand. "null" and "undefined" having different meaning is confusing. It's unclear whenonlogin
fires and frustrating it does not always imply user interaction. - The
watch()
API forces people to do unnatural things. The specific complaint was pulling emails out of sessions and embedding them in pages required more design change than desired. It should be much easier to build auto-login. - It's hard to manage multiple identities with Persona. It is too easy to end up with two persona accounts.
- My last used email address at a site is not remembered well. Logging out of persona eliminates the record of what emails you use where, and that is annoying.
- It is hard to share multiple email addresses. This can occur when a with a site wants to you allow you to register multiple email addresses, or set a backup address.
- Logging out of a site does not really log you out. Because you stay authenticated to persona, user expectation is defied when logging back in requires no credentials.
- There's no way to force re-authentication. "For any site that deals in information the user cares about, It's a requirement that you can force re-authentication to log in".
- This is not me, is not clear. When "This is not me" is used to switch between two accounts I own, the button language is wrong.
- Communication with the community needs to be better. Better meeting notes with more context, better distilation of current priorities. Better messaging of the things we're working on and their relative resourcing. Also, the community meeting wasn't announced as prominently as it should have been.
- Allow sign into one site to sign you into many. This is to support a more proper SSO solution for mid-sized businesses and public sites.
- Announcements of breaking changes are not prominent enough. For the community, Persona contracts and APIs sometimes seem to change without warning.
- It's hard to debug when persona breaks. (a sentiment echoed strongly by IdP implementors).
- Users can't understand that login means signup.
- Users don't understand the value of Persona. Sometimes it seems like facebook connect is just an easier solution. There is not sufficiently clear messaging to explain to users why they should share their email and set up what seems like another account.
- The popup would be better as a drop-down.
- Seems to not always work right on iOS.
- Persona is confusing for users on shared computers.
- It's not possible to use a shared email address. This is specifically when there is an email address that many people share.
- We want more site branding inside the dialog. When a user trusts a site but not persona, it would be better if the site could be more prominently branded inside the dialog to evoke user trust.
- Case sensitivity issues. Behavior is not always consistent when user's sign in with an email typed in mmixed case.
- IdP development is hard. Templates for an awesome authentication page that works across browsers and devices would help.