Nostr Web Client

I suppose NIP-05 syntax has been derived from LN addresses. I don't know if choosing an email like format has been the best idea (in both cases) because it's easily recognizable, but the goal is completely different and this can create a lot of troubles in the UX.

Why not simply change the separator and create a new and unique Nostr address?

user#domain.tld

If I see a @ linked address on a web page I expect to click it to open my email client. If I found a # linked address I know it will open my Nostr DM micro app to the user chat, or will fallback to the user profile.

In fact it can actually be a landing page pointer too; root domains can be ##domain.tld

Nostr apps that extensively use NIP-05, like the just proposed "shared login" by nostr:npub1l2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqutajft could improve the communication and avoid a lot of confusion.

Probably we can save newcomers from silly frustrations, e.g. "why I cannot login in Coracle with my Gmail?!", and perhaps prevent some security related problems too.

Maybe it can also help to spread the protocol, becoming a neutral brand in business cards and advertising.

Make sense?

Any cons?

#nostrdesign

Reply to this note

Please Login to reply.

Discussion

Gigi 2y ago

People see "#" and think #hashtag - but I agree with the criticism of using email-style addresses for NIP-05 (and lightning addresses, for that matter).

daniele 2y ago

I agree that the hashtag is not optimal, but unluckly the keyboard ascii subset is limited, and we need a characters easy to type.

Maybe with the user prefix the perception is different.

The alternative is a double char, e.g.

user!!domain.tld

signal_and_rage 2y ago

daniele 2y ago

Nice, but qwerty USA keyboard doesn't have it.

signal_and_rage 2y ago

does Shift + Right Alt + S work?

daniele 2y ago

Not exactly an easy combination to type :)

fiatjaf 2y ago

Mastodon also uses the @.

daniele 2y ago

Right, more chaos.

fiatjaf 2y ago

No, it's a good thing. The @ notation was never meant to be used just by email, email was just the protocol that was more successful. It's time to change that. Also, email is horribly bad, centralized, awful, no one should even be using it. We can't just let them win.

fiatjaf 2y ago

About Gmail, we can probably easily implement something like this: https://www.onebigfluke.com/2013/06/bootstrapping-webfinger-with-webfist.html

And it wouldn't be disruptive to NIP-05 at all as we would hardcode just the top 5 big email providers (I don't even know if there are this many).

daniele 2y ago

I knew it would end this way, capture the @ haha

Okay!

James A Lewis 2y ago

I feel the "at" symbol does make sense as an addressee within an address. It has multiple uses, but the fact that you're targeting someone within someplace is still true.

daniele 2y ago

'at' makes sense, in fact it has been a great pick for the email protocol.

The problem is actually the "multiple uses" with completely different goals. Try to explain to a casual user that your are giving him an email that does not work as an email, paired with a 65 chars password that cannot change... panic.

Should a Nostr identifier suffers this bootstrap weight? Or is better to start telling a new story with new rules?

I suppose that for this reason someone (maybe Alby?) started to propose to prefix the LN addresses with the emoji ⚡, to mitigate the problem and make them recognizable, by humans and programmatically.

James A Lewis 2y ago

I believe this is downstream of the domain name paradigm. We cannot get around the fact that our current system is set up with this user@domainname.tld system, not just for email but all sorts of protocols. User access and system login came before user-to-user-communication. Email is just the one protocol of which people think first when considering the "at" symbol. My point is @ is already very multi-use and is intuitive.

daniele 2y ago

You are right, the 'at' paradigm is present elsewhere, and it is always connected to the fact that the user is included, with his data, in that domain.tld system.

But Nostr is decentralized, the user identity and data are disconnected from the NIP-05 domain, this is just a proxy to the npub and some additional informations. Actually, I can have more NIP-05 too.

I don't know, I understand we are exiting a comfort zone, but I think it's worth thinking about now.

James A Lewis 2y ago

Right, we are trying to bridge from a user-domain paradigm to a un-domained key-pair paradigm. I don't know if a third way exists and would make for a reasonable and useful bridge.

NIP-05 is really just human readable pseudonym for an npub, by definition. I don't see a way to shorten or make human readable an npub without pseudonymization, so if already pseudonymous, what system would be more intuitive?

Niel Liesmons 2y ago

Yup, thought about that too. Very confusing indeed, but a # might be even weirder.

Is this an option? 👉 username(n)domain.com

Branding-wise that'd be something I can work with 😉

Thanks for making me think about the "neutral brand" aspect btw, there's definitely advantages there.

Matt Lorentz 2y ago

Matrix references rooms like this. While we’re at it let’s allow people to have multiple NIP-05s.

In fact now that nostr:npub1xs643urmqllukf9j0decztt56nlc73hgr65s8u087d7npkg8hn7qy2m204 is speccing out a signin with username and password thing maybe this sort of username should just have a new NIP.

daniele 2y ago

I fell the problem.

But the double @ is confusing, imho.

If we have to stick with @, let's capture the user@host syntax.

I was talking about this matter here:

nostr:nevent1qqsy282x04w20mcs2h2jd7eeltxj23r54enxwrhhc335rmcnafwe5qqppamhxue69uhkummnw3ezumt0d5pzq77777lz9hvwt86xqrsyf2jn588ewk5aclf8mavr80rhmduy5kq9qvzqqqqqqyef2x9e