Relay runners: Which relay implementation are you running? What is your experience with this software?

Shoes are shitcoins

True #[48], we need a section for the art builders

💯

If this pace keeps up, we'll be unstoppable.

But if everyone has the same/similar relay arrays it’s an easy attack vector, no? Unique relay arrays don’t offer such good attack vector

Not sure if you have the same thing in mind I do, but i was thinking the logic can go like this:

1. you search for a user

2. client tries to get their info from your existing relays

3. if you don’t have any relays in common, your client can then use one of the big relays currently used as “defaults” as a bootstrap to look up which relays that profile does use

4. your client then connects to one or two of those relays at random so that you can reliably see their posts

Which metric should decide when redundancy is needed?

I don't want a "partner" for those walks because I just need to be alone and think why I allow myself to be depressed. If something distracts me the whole walking will loose its purpose. I use the music to isolate myself from the surrounding noises.

I do the same and I force myself to go for a walk.

Obviously yes... manual or automatic relay management can coexist.

Yes agree. However, I think the option for manual management should still be there, just to endure censorship resistance. If you take the choice away of choosing your relays manually means that controlling your relays = controlling you. Even so most people won’t use it anyway. Or do I miss something?

Of course, new users would have to bootstrap somehow in a more centralized way, I guess. Thing is, relay subscription could evolve in the background without the user noticing. Just making sure to keep in touch with the most people with the least possible amount of subscriptions.

I’d still keep a list of relays, so that the client can start quick, without needing to download a ton of profiles doing so.

Thats great but not applicable for new users right, since they don’t have a following list yet, right?

The concept should also include the follower list, to make sure you reach your audience or is it included in the same list?

Dear devs,

I think how clients set the default relays is a very important key topic to ensure future decentralization of the nostr protocol. The non-tech masses won't care about relays, they just want it to work. Max performance/connectivity and max decentralization are not the same route. We have to find the right balance.

We should develop a best pratice how to set default relays on clients.

Some form of randomness is vital to not fall for the centralization trap. It should be combined with some important metrics (eg traffic, reputation, ping, uptime, location ...). Defining these metrics and combine it with randomness is the key from my perspective.

Example for 12 default relays:

1. Connect to random ~3 relays with high user count, reputation score > 90 and uptime >99

2. Connect to random ~4 relay with mid user count, reputation score > 80, ping < 500 and uptime >95

3. Connect to random ~5 relays with low user count, reputation score > 70, ping < 100 and uptime >90

Let me know your thoughts on this.

Adding #[18] doing a Nostr lib.

But also tempting to keep refining your code when it's already just good enough!

😂😂 hey at least you realized!