I am just starting to realize that this idea of general relay lists doesn't make any sense.
Home feed -> relay set from follows
Global feed -> specific relay or relay set
Chats -> each has their own master relay
Private Messages -> shared relay between each pair of two users.
Notifications feed -> all relays in existence
User profile -> all relays in existence
Event info lookup -> all relays in existence
Login -> users preferred/home relay set
I think that makes more sense.
If clients are implementing NIP-65 properly than notifications do not need to listen to ALL relays
Good thinking
But how do you get the follows relay list, if they themselves don't have a relay list?
Basically the only relay set we need is where each user is posting. Each chat should have a relay set registered. The app then needs to be able to do a wide call in all 1000s relays to figure out where things are.
What I like to see is a #nostr Client where you can create your own lists of people you follow (with the right relays ‘attached’). This way a user can quickly switch to a list where he can see all posts regarding an event (e.g. nostrica in my example).
I started to design this idea over the weekend:
Topics should not be tied to People. Picturing a topic by a list of follows is not ideal. A better model is Slack's workspaces, where the user mentally shifts into a community/topic. That can be achieved with a relay set switch for read and write. In that way, you dont get your Bitcoin follows posting about their diets.
Hey Vitor can the user set the recommended relay or is it up to the app?
Right now it can, but besides the home relay set of the user, there might not be a need for any other relay setup.
?cid=2154d3d7129894d7e3a9993e51e6c322ca32b83b4036dfeb&rid=giphy.gif&ct=g
#[2] maybe inspirational for you. Sound in general like a good plan.
Yeah indeed it was originated from a conversation we just had on a thread about the topic 😄
100%. I find myself gravitating toward clients that give you more control over the relays. They're very important to use deliberately and understand, even for new users..
It's going to be important to understand them, nostr is not just a single pane of glass that can blast a message to the entire world. That doesn't scale.
Have you looked how gossip does it?
How would a brand new user know what relays are best? Which are paid and so forth... all this information should be provided within the client surely...
⚡⚡
#[2]
We are moving towards the Gossip model and perhaps Areas of interest...
#[0]
👀 keeping a close eye on how this plays out.
After listening to the "uncle bob" episode of nostrovia i deff agree with a return to bbs style topics.
#[0] has been trying to communicate this for some time now 🙂
Client apps need smarter “which relay” algos for inbound and outbound.
#[1]
Do it👹😇
Nostr success may depend on how well relays work out of the box for non tech savvy newcomers
Also things like, reactions/zaps should be loaded from a wider range if relays. Also, if someone I follow replies to a note, i want to be able to see that note, even if its not on a relay im connected to
I know you’re looking for a client side fix but have you tried https://filter.nostr.wine in the interim? Will help expand your note coverage in situations like this without you having to add 10+ additional relays.
Oh thats awesome! Great idea
mark
For the “all relays in existence” cases, would you say the most efficient way to handle those is to use a relay indexing service like nostr.band?
#[4]
Yes! Relay management needs to be much more granular, maybe automated when possible.
That does make more sense.
这不知道哇，有同好的，也许他们就免费帮你搞出来功能了。
Amethyst 的开发者有说过他的基本思考的，你看看。
#[3]
#[0]
How did this fare after two months, more NIPs, and sort of new relays being built and implemented?
