Skip to main content


#TIL about the Join Jabber collective:

https://joinjabber.org/about/goals/

Their aim is to make the federation of chat servers using the XMPP protocol accessible to more people, and improve #UX across the whole network.

#chat #federation #jabber #XMPP

in reply to Strypey

Is there mandatory end to end encryption yet? It's a legacy protocol as far as I'm concerned unless/until we have that.
in reply to Charles U. Farley

> Is there mandatory end to end encryption [in XMPP] yet?

Nothing is mandatory in XMPP except for the core and that's not going to change. But perhaps I'm taking you too literally? If your question is do all (non-beta) clients support E2EE, I'm pretty sure that's a goal. I believe there's been some funding grants handed out to client devs to help them finish/ audit OMEMO implementations.

Care to correct/ add to any of this @joinjabber?

in reply to Strypey

Indeed not mandatory, but there are some XMPP clients that offer the option to make e2ee the default in private chats. In general this website still gives a good overview on the state of e2ee support in XMPP clients: https://omemo.top/

There is also ongoing work in some clients to implement a newer version of OMEMO (nick-named newmemo or omemo2) that offers much improved metadata encryption compared to the current standard that encrypts the text body. #xmpp #jabber

in reply to JoinJabber

> a newer version of OMEMO ... offers much improved metadata encryption compared to the current standard

Good to know! I presume Conversations is one of the clients working on this? If so, it will available to people using @snikket_im as soon as its stable.

@freakazoid

This entry was edited (1 year ago)
in reply to Strypey

As long as people have to take action to enable encryption, it will be possible to make the case that the use of encryption is evidence of wrongdoing. I think there is a big opportunity here for XMPP to significantly move the needle on privacy because, unlike Signal, it's not "primarily" about encryption.

Right now I use Signal with my friends and family. I don't see moving to Jabber unless/until the iOS experience improves (very slow/missed messages due to battery optimizations) and encryption becomes the default on some usable client for each of the platforms used by my friends, mostly iOS and Android.

in reply to Charles U. Farley

That has already happened 😀

Conversations and cheogram has a setting to make all conversations encrypted by default.
Monal also on IOS. Since version 5.3
And monal is also on mac.

in reply to MSavoritias

By encrypted i mean e2e with omemo.
Xmpp has had encryption on everything (tls) for a long time.
in reply to MSavoritias

And conversations, cheogram have a setting that doesnt allow you to disable omemo at all.
Not sure about monal.
in reply to MSavoritias

Not a setting to make it the default. It needs to be the default out of the box, both so I don't need to give family and friends additional instructions, and so that it cannot be considered evidence of wrongdoing.
in reply to Charles U. Farley

Yeah i get it.
Signal has an easier way because its centralized and non democratic. What signal corp says goes.
With #xmpp we go about things the standards and democracy way. So hopefully in a bit we can have encryption by default. PGP is already almost phased out as completely insecure compared to omemo.

So what i can offer now is that its just a setting that can be enabled. Which i have done for friends and family. Because they are going to ask for help to set it up anyway 😀

in reply to MSavoritias

Yeah decentralized and democratic is definitely my preference. I will take Moxie as a benevolent dictator, but dictatorships are only ever benevolent temporarily.

Plus, while I totally understand the reasoning, I don't like using my phone number as an identifier.

in reply to Charles U. Farley

Moxie is no longer involved right?

Also remember even though Signal is e2e by default, they forced people to enter a trivially breakable 4 digit PIN which caused your keys+history to be sent to Google's servers, and only walked back the "force" part after most of the damage had been done due to backlash.

This can't happen in an open protocol with multiple independent implementations, pick one with e2e by default and be happy.

If you are actually serious about privacy you can even run your own server so no one has your metadata, not even Signal.

in reply to moparisthebest

I haven't been paying as much attention as I should. I think I probably heard Moxie was no longer involved at some point and promptly forgot.

If you're talking about their "secure value recovery" thing, I can understand why some people would be concerned about it, but in general I'm not concerned about encrypted data being stored on servers. Their secure enclave method seems very similar to what Apple said they were going to do for secure backups and then quietly didn't.

I generally compare anything like this to the available alternatives rather than to some absolute ideal. If there were an alternative an alternative available that I felt was overall better and that I could get friends and family to use, I would switch.

in reply to Charles U. Farley

As for why it's important to me that encryption be the default:

"Although the exception purports to protect online platforms from liability for offering encrypted services, it specifically allows the use of encryption to be introduced as evidence of the facilitation of illegal material."

https://www.eff.org/deeplinks/2023/05/stop-csam-act-improved-still-problematic

Optional encryption might as well be no encryption if it can bring the rubber hose down on you. The only protection from rubber hoses is for the people with the rubber hoses not to know who to use them on.

in reply to Charles U. Farley

> Optional encryption might as well be no encryption if it can bring the rubber hose down on you

FWIW Nothing in this article suggests this law applies to people using a third-party service. Only to the service provider.

@moparisthebest @joinjabber @msavoritias

in reply to Strypey

I guess I don't understand how that works. Someone on the platform uses encryption and a prosecutor can use that as evidence the platform is facilitating that person's transfer of illegal material without also using that against the person themselves?

Even if something like that is the case, it's still an example of people assuming that taking positive action to use encryption indicates possible wrongdoing, whether or not it can be presented as evidence against the individual in court today.

in reply to Charles U. Farley

> Someone on the platform uses encryption and a prosecutor can use that as evidence the platform is facilitating that person's transfer of illegal material without also using that against the person themselves?

IANAL. But my rough understanding is that this digitally illiterate bill;

a) only targets service providers (not users) by making holes in Section 230 (Safe Harbour for platforms hosting third-party content with moderation).

(1/3)

@moparisthebest @joinjabber @msavoritias

This entry was edited (1 year ago)
in reply to Strypey

IANAL. But my rough understanding is that this digitally illiterate bill;

b) whether the E2EE is on by default or not makes no difference. If the service offers it, this bill would make them potentially liable for accusations of use involving CSAM.

(2/3)

@moparisthebest @joinjabber @msavoritias

This entry was edited (1 year ago)
in reply to Strypey

> taking positive action to use encryption indicates possible wrongdoing

If people have fallen for this wrong-headed notion, you've got a more basic problem than legal risk from bills that will hopefully never pass.

Ask them if drawing the curtains before getting undressed indicates possible wrongdoing. Turning on encryption, just like choosing a service that uses it by default, is just the digital equivalent.

(3/3)

@moparisthebest @joinjabber @msavoritias

in reply to Strypey

I don't think any of that changes the desirability of making encryption the default by default. And maybe not optional at all.
in reply to Charles U. Farley

have you thought about how this affects our historians? If society exists long anufe for this to matter then we are creating problems for future generations.
in reply to vagabond

@Hamishcampbell @moparisthebest @joinjabber @msavoritias I have thought about how this affects historians. Same with auto-deletion of old posts from the Fediverse. But the larger impact is whether historians will be recording the victory or failure of fascism.

If we really care about history, the correct approach is to create archives which are inaccessible for some specified period of time. That's hard to accomplish technologically, but could potentially be implemented within some trustworthy jurisdiction.

The archive itself could use some combination of a time lock and dead person switch. If someone tries to forcibly access the archive, it gets wiped.

in reply to vagabond

@Hamishcampbell
> have you thought about how this affects our historians?

1) If data is important for posterity, it needs to be published, unencrypted, in an easily digestible way (eg blogs, videos), using properly documented open formats. Not squirreled away in encrypted chat logs.

2) Whether data is archived at all, and using properly documented formats, is a much bigger problem for future historians than whether it's encrypted.

(1/2)

@freakazoid @moparisthebest @msavoritias

in reply to Strypey

@Hamishcampbell

3) If encrypted data is archived for long enough, university history departments will eventually have computers powerful enough to decrypt it by brute force. By which time anyone involved in producing the data will most likely be long dead.

(2/2)

@freakazoid @moparisthebest @msavoritias

in reply to Strypey

@Hamishcampbell @moparisthebest @msavoritias This is not necessarily true, because of the exponential increase in the size of the search space with the number of bits. A computer the size of the Earth trying one key per nanosecond per atom would still take the age of the universe to crack a 256 bit key by brute force.
in reply to Charles U. Farley

@freakazoid
> A computer the size of the Earth trying one key per nanosecond per atom would still take the age of the universe to crack a 256 bit key by brute force

I'm just very aware that we don't know what we don't know. No matter how smart we are. Remember Bill Gates' infamous claim that nobody will ever need more than 64KB or RAM? People quote that as evidence that Gates' technical knowledge was weak, but that's clearly not the explanation.

@Hamishcampbell @moparisthebest @msavoritias

in reply to Strypey

@Hamishcampbell @moparisthebest @msavoritias I think Bill Gates was suffering not from a lack of knowledge but from a lack of imagination. Which, when money is limited (as it hasn't been during each of the tech bubbles including the one that's deflating right now) is the fundamental limiting factor in entrepreneurship of any kind, capitalistic or otherwise.
in reply to Charles U. Farley

@Hamishcampbell @moparisthebest @msavoritias As for what we don't know, differential cryptanalysis was a pretty huge leap, and I imagine we'll be able to use machine learning to automatically find "shortcuts" in existing algorithms and implementation. Kind of like how AlphaFold managed to advance the state of the art in protein folding by an order of magnitude.

Of course, once we're applying AI to cryptology, I imagine we'll apply it to developing new algorithms in an "adversarial" manner as well. They may start lasting for much shorter periods of time, though, depending on how progress on AI goes.

in reply to Charles U. Farley

@freakazoid @moparisthebest @msavoritias

something you are not addressing, why do we need encryption at all in the #Fediverse it's all currently #4opens and is very successful because of this - why do we need encrypted DM's that a current chat app cannot do?

in reply to vagabond

@Hamishcampbell
> why do we need encryption at all in the fediverse

As @freakazoid said, that isn't what this thread is about. We're talking here about chat apps. But since you asked...

The most obvious reason is to protect the privacy of people exchanging private posts. But with the bonus that if all posts considered private were encrypted, then all unencrypted posts could be considered public, archived, indexed etc.
@moparisthebest @msavoritias

Lo, thar be cookies on this site to keep track of your login. By clicking 'okay', you are CONSENTING to this.