Because edits only exist in one client, it makes it possible to target that client using the protocol. To stop the attacks, you can 1. drop edit support, since it's not interoperable, 2. convince other people to support edits so that the attacks against amethyst become attacks against all complying clients (or 3. find some other clever way to defend).

Instead of resorting to character assassinations maybe we should take this opportunity to learn about how protocol attacks work, and improve on building consensus.

Reply to this note

Please Login to reply.

Discussion

Or 3. People realize that even with the attacks people are still using it and grow the fuck up.

Whatever happened to "permissionless decentralized network"? If you (we) can't deal with even simple attacks like this then nostr is DOA. The main defense to protocol attacks is counter-attacks, either via client feature support (or lack of it), or specially crafted events. The result in either case is that users get a broken experience if they or people they follow use "bad" features.

The edit spam is the latter kind of attack. If users want edits but don't want spam, they can keep using Amethyst and stop following fiatjaf. If they want to keep following fiatjaf but avoid spam, they can use a new client. Be glad I guess that the spam is coming from just a few users, not from an entire hostile client like might someday arrive when Meta finds nostr.

> Whatever happened to "permissionless decentralized network"?

You tell me. You are trying to tell not only me as dev but all my users what they can and cannot use.

Amethyst is working great even with the attacks. And yes, many people are blocking those that are trying these attacks.

Most importantly, everybody is learning who is here to impose their will on users and who is here to figure out what users want to do and support them.

So why are you so bent out of shape about it? Is there no reasonable motive for doing what fiatjaf is doing that doesn't depend on him acting in bad faith?

There was a reasonable motive that has been debunked over and over again. Now it's just recycling old arguments that don't make any sense. To me, that is the definition of bad faith. He is not interested in leaving people alone in the way they want to use Nostr for.

"debunked" Who exactly is the arbiter of truth here?

No one told you you "can't" implement edits, just that you "shouldn't". That's an appeal to reason, not an imposition of power. Edit spam is a demonstration, within the rules of the protocol, of that disagreement. Now, if fiatjaf deleted your PR from the github that would be bad, but nothing like that is happening.

> "debunked" Who exactly is the arbiter of truth here?

Users. If they want it, they will have it.

That's not debunking, that's user preference.

Sure... Never fight user preference. Or you are going to lose.

Always listen to your users. Never take what they say at face value.

can you link to what was debunked? I don't think I've seen your counterargument

I think you'll have to build snopestr first

I don't think most Amethyst users care that you guys are spamming your relays with edits.

You're not attacking Amethyst.

Was it a war? I enjoyed watching

You know its not about control, Vitor, come on!

Edits are an imposition in all clients that centralizes the protocol.

Why? Are you imposing kind 20 feeds on other clients with Olas? I wouldn't think you do.

What's wrong with a feature people want? If other clients don't support it, they don't support it. I can give you a long list of things clients don't support, and that I would consider essential. Edits is the least important of these things.

Edits overload kind 1s, creating conflicts in what people might see in different clients when viewing the "same" event. It's quite different from purely additive features.

Deletions overwrite kind 1 too. Nobody cared for months. Some clients still don't.

Deletions are either effective (the note is gone), or additive (the note is "deleted"). They don't change anything in place. The effect of a successful delete is the same as not being able to find a note, which has to be handled in any case.

You will probably still find it on njump, except they changed something recently.

Yeah, it was there a while ago when I originally did this experiment, and I think it was around in people's local Damus cache.

Maybe one day you will also see how much edits just work really well.

It could happen! And maybe someday you'll join the anti-edit masonic cabal.

The option will always be there.

If people agree with you, they can just not use the feature.

I don't need to make that call. My users are smart enough to make that decision for themselves.

Anti-edit masons 😂 that'll be the day.

Edits do work really well in theory. But attackers cause that theory to be reworked.