this post was submitted on 10 Oct 2023

24 points (72.2% liked)

Lemmy

2172 readers

106 users here now

Everything about Lemmy; bugs, gripes, praises, and advocacy.

For discussion about the lemmy.ml instance, go to [email protected].

founded 4 years ago

MODERATORS

[email protected]

Could a "login with lemmy" replace all the login with Google/Facebook we see everywhere? (lemmy.fbmac.net)

submitted 1 year ago by [email protected] to c/[email protected]

18 comments fedilink hide all child comments

I love the convenience of not having to create a password everywhere I need to be authenticated. It would be interesting to be able to use lemmy instead of feeding more information to these big corporations.

all 20 comments

sorted by: hot top controversial new old

[–] [email protected] 33 points 1 year ago (1 children)

The short answer is: No.

Each site would need to implement this feature themselves. Think of all the potential platforms out there: lemmy, reddit, pinterest, instagram, linkedin, tiktok, snapchat, tumblr, whatsapp, friendster, quora... this list is long. Like REAL long, and Lemmy is nowhere near the top in popularity. Probably not even in the top 50 yet. Why would site developers implement Lemmy sign on over any of these others?

We did have a concerted effort at some point to adopt openid which was specifically designed to unify all these signons in an open manner, but instead we ran into the Standards Problem: xkcd 927.

[–] [email protected] 19 points 1 year ago (1 children)

To add to what others are saying... you really don't want or need this anyways. Once passkeys are used broadly or well supported by password managers, logins will be pretty low friction.

[–] [email protected] 2 points 1 year ago (1 children)

I've seen passkey added to 1password, do you have an eli5ish way of describing that? Is it like ssh keys?

[–] [email protected] 6 points 1 year ago (1 children)

If you're familiar with SSH keys, it's basically the same thing but written for everything to use rather than just written for SSH.

Passkeys can also be managed by TPMs/device hardware to provide private keys that (assuming things are properly designed) can't be stolen and are bound to the device itself. This isn't a requirement however, which is why things like 1password, Chrome, etc can sync passkeys.

They're basically syncing the one-or-more private keys that back your logins.

[–] [email protected] 1 points 1 year ago

Very neat! I'll have to try them out on a couple sites

[–] [email protected] 17 points 1 year ago* (last edited 1 year ago) (1 children)

No, because Lemmy doesn't support Oauth2 yet. And even if it would support it, at most it could be a "login with lemmy.ml" or similar instance specific button as the protocol requires a specific endpoint.

Edit: see other comment in this thread. With the OIDC discovery extension to OAuth2 it might be possible, but I haven't seen that feature being used this way in the wild yet.

[–] [email protected] 3 points 1 year ago (1 children)

when you type [email protected] I already know what instance you're from

[–] [email protected] 8 points 1 year ago* (last edited 1 year ago) (2 children)

Yes, but that is not how Oauth2/OIDC works (the old OpenID did, but it has been largely abandoned).

One of the reason this approach was abandoned is that these external login automations are very easily abused for spam if you allow arbitrary instances as the auth endpoint.

[–] [email protected] 3 points 1 year ago (1 children)

So this is why we've been seeing rows of "Login with $SpecificProvider" instead of a universal format using username@provider as we all hoped?

[–] [email protected] 5 points 1 year ago* (last edited 1 year ago)

The old OpenID didn't see much uptake (because of the spam issue) and the alternative Oauth2 that was AFAIK mostly pushed by Google is clearly designed for the purpose of large centralized providers. So I don't think there is a direct causality, but yes it is related.

Never the less Oauth2/OIDC works quite well and is clearly better that most of the alternatives still commonly in use.

[–] [email protected] 0 points 1 year ago* (last edited 1 year ago)

you could accept logins only from instances that have enough trust on fediseer, I think this would work better than the old openid

[–] [email protected] 13 points 1 year ago

There’s no way I’m letting a lemmy instance hold the keys to any of my other accounts.

[–] [email protected] 11 points 1 year ago (1 children)

I feel like logging in with Facebook and Google on sites is just as much about them gaining access to scrape more info about you as it is for your "convenience." While there could be value with Lemmy, it's not nearly what it is with Facebook and Google. So I would say not likely.

[–] [email protected] 7 points 1 year ago (1 children)

It absolutely is, anytime you see a login with Facebook it is about tracking you. Even if you are Not logged in, and you do Not use the button, the Facebook Pixel is still there tracking you.

[–] [email protected] 2 points 1 year ago

Hence why I block it.

[–] [email protected] 6 points 1 year ago* (last edited 1 year ago)

I ran into this issue when writing the Canvas authentication thing

There is a standard to simplify authentication and identification across the entire web but it isn’t fully implemented everywhere

OAuth2 is the big name in this, it supplies authorized requests to access data from other companies and services

OpenID is related to OAuth2 but it only supplies identification in a standardized way. OpenID has mechanisms to announce that a specific domain has support for this and how to automatically register for it (removing the need to have a bunch of login buttons)

For Canvas 2024 I’m implementing drop in implementations for popular Fediverse software (including Lemmy)

Hopefully more fedi software implements OAuth2 or at least OpenID to vastly simplify authentication (and possibly replace “login with google”)