The problem with outbox is that the user has the mental burden now that not only should he understand the importance and implications of cryptographic keys but also do some research on relays in a similar fashion.
We could say this is the end-goal to normalize concepts like keys and relays but there has to be a learning path leading to that goal, coupled with the right incentives and UX.
This path is lacking from each and every nostr client out there right now. IMO it has to be communities bridging the gap, because you have an already bootstrapped natural trust that you take to the digital realm with you.
We cannot assume this trust will be built effectively by automatically following ppl, or attaching a default social graph to the user. This is the mistake of big tech. They want you to be comfortable, without friction so they control your experience on all levels.
If we want a realistic path forward we need to onboard real communities to nostr and create software that effectively builds up this learning path towards increasing user-control.
This is a tremendous task compared to go full haywire with outbox right away, or the other end of the spectrum, hardcoded relays forever.
It takes time but I'm hopeful we'll get there, we just need more focus on real communities.
Replies (8)
Isn't this where starter packs can play a role? Or use the relays that the person who referred you uses?
I think we have mistaken the Protocol for automatic Community building.
We fell in love with the tech concepts and forgot the human percepts.
We Robot! 🤖 😆
I'd say outbox decreases such friction, because you don't run into these edgecases where people mis stuff because they did not specify some relay. The whole point of outbox is that it does not matter on what relay events are stored.
In relation to the community thing:
I agree, but also note Nostr got its innitial bootstrap from the bitcoin community. This is/was great because of the type of people (tech savy early addopter people that understand keys, or atleast some of those elements), that allowed for Nostr in the broad sense to develop.
And given that everyone is probably not just a Bitcoiner, but also something else, they serve as anchor points for more specific 'community' stuff. Its a natural progression if you ask me. The more bitcoiners we onboard, the more demand for niche things with tailored clients to facilitate them will rise. The relay based groups ecosystem made great improvements the last period, where they sort of work, with more than 1 client that is decent.
A 2-party onboarding process might be useful.
It could essentially clone and initialize a new account to match sone aspects of a purple-pillers existing account (relays, maybe follows or other stuff that the purple piller can "push" to the new user's initialization process).
I'm thinking something like a SimpleX pairing process, but with more interactivity between the inviter & the invitee.
Communities are living starter packs.
Yes bitcoiners can help, but I wonder how many real bitcoiners there are on nostr, who have the skills and the time to onboard others. Very few I believe.
And even if that's true, it is a daunting challenge to make outbox really work, and the inevitable rough edges make user retention hard.
It is telling that apps with plenty at stake, e.g. Primal and Satlantis don't even connect to relays directly. Damus has not used outbox for a long time I don't know about nowadays.
Guess why?
When you want to publish a product that works, it is quite important for stuff to load.
If you hardcode relays or use some caching server shenanigan with your own api things will definitely load using the same app on both devices.
This is not the end goal but you can onboard ppl, things are still signed etc.
I disagree with specifically the proprietary api method but it showcases the problem:
We are told "outbox is an art so we don't make a spec out of it". Okay, but this is sort of a cup-out.
Outbox is a brand new paradigm to build apps. No one has proven to my knowledge that it works in production with more than a handful of users. How many web apps use 100s or thousands of server lists to calculate the right subset **for different contexts the user can navigate to in the app?** and you have to score relays that misbehave and handle disconnections, auth, timeouts, and resource management to not exhaust battery and data of the user rapidly. To name a few.
What if we take this to the labs and create sth we can use in the meantime that also showcases the fantastic properties if nostr?
Rather than chanting "outbox" on no-true-scotsmen ?
Communities are a better solution to these issues bc most of them will start out on one relay. If we make these compatible with each other, what we get is a Fediverse that is 100X better already, with opportunities for further development.
I love bitcoiners but outbox was born out of the bitcoiner mindset that diverted our attention prematurely I think.
Lo viste?
nostr:nevent1qqsrf44kha3q88gcrkquvstlxhmnyesj32623n3pl4megm7652ycwkcprpmhxue69uhhyetvv9ujuen0w4h8gctfdchxvmf0qgszp55czrt2t7ftq3dduqhthtwfqdkhg8xxs6cqg9wy9dprdlj26tcrqsqqqqqphduvu7
