Live data from Hacker News

Open Letter to Mozilla: Bring Back Persona

stavros.io

171–180 of 243 posts

Re: Open Letter to Mozilla: Bring Back Persona

#171
post #158

Earlier quoted context omitted.

"Why not just...?" I enjoy that you offer up a quite complex, confusing, often insecure, and error-prone methodology with this phrase.

How is sending a user an email complex, confusing or error-prone? As for security, given that an account can already be accessed if one has access to the holder's email (which is the case for any password reset flow relying on email), this is no more insecure than the current state. Seriously, this sort of thing is not complex, not at all .

Complex situations:

- I'm logging in from a new computer. Maybe a computer I don't trust enough to log in to my email (which controls nearly everything, as you note). Not all services are that important or need the same level of care as email.

- I'm a user that has greylisting enabled on my mail server, potentially delaying the email for several minutes.

Confusing:

- I'm a user that has never seen/used this authentication workflow. While user/pass is annoying, it is widely understood. Sometimes "ease of use" is really just "what I'm used to". Thanks to facebook and Google authentication, the "Login with..." button is now "what I'm used to". I've never seen the authentication method you've described in 20 years of web use, except as a password reset method.

Error-prone:

- Implementation would have to be done by me on all of my sites. There is no off-the-shelf plugin for this for Drupal or Wordpress or Mediawiki, which run 90% of the websites I am responsible for. I could develop it but I am certain I would make mistakes in said implementation. I should be clear that I think any form of authentication is error-prone, unless it's already written and well-tested. User/pass and optional email is already written and well-tested for every site I maintain.

Security:

- It forces login to email to login to your site. This increases the risk of a user logging in from an untrusted computer or in an untrusted location to an incredible degree.

- Also, one can easily add security questions, or other second factor, to a password reset form, to make it more secure than your solution. Password resets are rare; happening every year or two, most likely. Logins are daily. One cannot impose those security questions on daily login form without paying the same annoyance price as user/pass (higher because it is unfamiliar). Cookies can mitigate this somewhat, but it doesn't solve multiple devices, and security-conscious environments that don't maintain cookies on logout.

If you believe this is an acceptable workflow for authentication, I won't try to stop you. But, it simply isn't a solution for my sites. I don't like user/pass, and I want something better, but I don't believe what you've described is something better. It's a neat idea, but not a solution to logins.

Re: Open Letter to Mozilla: Bring Back Persona

#173
While I think Persona is way better than password-based auth, I can't help but feel that it's also a sideways step. No, authentication should be provided by the user-agent, not by some third party identity provider (even if you control it by running your own). We have this for ssh: you have your private ssh key and it's up to you to manage having it in the right places (work desktop, laptop, phone, etc.) It's clunky, but that's OK: it's aimed at tech people.

Browsers should do something similar: provide a storage for private and public keys as proof of identity. But they should additionally provide a way to sync those identities across multiple browsers and machines. This latter concept will be familiar to anyone who has used LastPass. The former, well it should basically be a dropdown with your identities, and you choose one before clicking the "Login" button.

Re: Open Letter to Mozilla: Bring Back Persona

#174
Persona is a neat idea, but there's a better one: get rid of password logins altogether. I should just be able to enter my email on ANY site, and they'll send me an email with a login button. I click the login button in the email, and am automatically logged into the site for as long as necessary.

It's exactly as secure as the "Forgot my password" reset nonsense, but streamlined to be way easier on the user. Then you only need one password: the one to your email. This removes the need for Persona to "add support" for different email providers. Your provider doesn't matter, just the fact that you can receive emails!

I'd really wish more sites would just do this anyway. For instance, every time I get an analytics daily report from Fabric.io, I click the link in the email, and it asks me to login. WHY ARE YOU ASKING ME TO LOGIN, WHEN I JUST CLICKED AN EMAIL YOU SENT ME? It's obviously me!

Re: Open Letter to Mozilla: Bring Back Persona

#175
post #66

Earlier quoted context omitted.

I don't see a problem, it just means you can login with multiple emails then. They're still unique to you. LinkedIn and Facebook and other services already do this. A username/real name can still be used as the "name" if this is an online community or something similar.

LinkedIn and Facebook can do this because your email is not your username. There likely isn't really a username in those systems, just a user id. Almost all systems have a user id, but the distinction here is slightly different. Instead of a user record which has an id, username and possibly full name, Facebook likely has a user record with an id, a (display) username, and then there is a separate set of auth records…

What you're describing is "use a synthetic primary key for your User", which is something even the most naive junior web developer does automatically these days. There's nothing special about LinkedIn and Facebook except that they went through the trouble of adding additional authentication mechanisms. You can too, no matter what you treat as a "username".

Re: Open Letter to Mozilla: Bring Back Persona

#176
post #82

Disclaimer: I work for Mozilla, I maintain django-browserid (and StravosK is a valued contributor many sites. This is all just my own personal opinion. I was very bullish on Persona early on, but the fact of the matter is, we failed . And not just because (as I feel is being implied) some higher up suddenly came over and asked for an unreasonable amount of adoption for a revolutionary product. We failed for a thousan…

I agree with you on most of this. And yes, a "Persona 2.0" should really just take the Persona lessons and learn from them, avoid the mistakes. That's also why I mentioned in my other reply that I'd be willing to help a project that has a chance to succeed - and that means being lead by someone who has themselves learned all the lessons from Persona. Someone who was on the original dev team. But this also needs backi…

> (there aren't many)

Because this is so important maybe it should be left to just one entity to run. Ideally a federated identity login thing should be backed by a plethora of orgs.

Let's make a list?

Redhat, Linux Foundation, Raspberry Pi, Ubuntu, Mozilla, LibreOffice, Eclipse, GNU, ... ?

Re: Open Letter to Mozilla: Bring Back Persona

#177

While I think Persona is way better than password-based auth, I can't help but feel that it's also a sideways step. No, authentication should be provided by the user-agent, not by some third party identity provider (even if you control it by running your own). We have this for ssh: you have your private ssh key and it's up to you to manage having it in the right places (work desktop, laptop, phone, etc.) It's clunky,…

To some extent, this is similar to SSL/TLS certificates that can be used to authenticate the user (not just the server). If I recall correctly, http://startssl.com/ does that.

It's not a very intuitive mode of authentication, but if the UI was improved, and combined with a sync service (& encrypted with a passphrase), I guess it could be usable? (also, presumably it requires the site to use https, but that's also much more expected today than it was 10 years ago)

Re: Open Letter to Mozilla: Bring Back Persona

#178
post #105
post #101

Earlier quoted context omitted.

"the annoyance that is "Sign up with a username, email address, and password"." I know of an easy way to remove 33% of that ... Email addresses are globally unique, and everyone already knows theirs ... email + password seems like a simple, well accepted and workable solution. Right ?

Because i'd love for my reddit username to be my email when i decide to play devils advocate, and campaign for Trump.

If you have that kind of site, then do what reddit did -- cut out the email address. Still a 33% saving.

Re: Open Letter to Mozilla: Bring Back Persona

#179
post #66

Earlier quoted context omitted.

I don't see a problem, it just means you can login with multiple emails then. They're still unique to you. LinkedIn and Facebook and other services already do this. A username/real name can still be used as the "name" if this is an online community or something similar.

LinkedIn and Facebook can do this because your email is not your username. There likely isn't really a username in those systems, just a user id. Almost all systems have a user id, but the distinction here is slightly different. Instead of a user record which has an id, username and possibly full name, Facebook likely has a user record with an id, a (display) username, and then there is a separate set of auth records…

You are describing the tripartite identity pattern: http://habitatchronicles.com/2008/10/the-tripartite-identity...

Re: Open Letter to Mozilla: Bring Back Persona

#180

Earlier quoted context omitted.

I agree with you on most of this. And yes, a "Persona 2.0" should really just take the Persona lessons and learn from them, avoid the mistakes. That's also why I mentioned in my other reply that I'd be willing to help a project that has a chance to succeed - and that means being lead by someone who has themselves learned all the lessons from Persona. Someone who was on the original dev team. But this also needs backi…

> (there aren't many) Because this is so important maybe it should be left to just one entity to run. Ideally a federated identity login thing should be backed by a plethora of orgs. Let's make a list? Redhat, Linux Foundation, Raspberry Pi, Ubuntu, Mozilla, LibreOffice, Eclipse, GNU, ... ?

I could see a joint effort, similar to what happened with Let's Encrypt. The players who benefit are different though so it's harder to get the same kind of pull.
Post reply on HN