if there is no unique identifier in NIP-B0 then how can we track it properly.

If we track only by kind:pubkey:d-tag, what happens when a kind receives comments and zaps, but the d tag is later updated? How can we track it again along with its related comments and zap kinds?

Additionally, if a bookmark is deleted after receiving some zaps and comments, and later the user creates a new bookmark with the same URL, will the previously deleted comments and zaps be mistakenly linked to the new bookmark, even though they are not actually related?

Reply to this note

Please Login to reply.

Discussion

if you edit an event the d_tag remains the same, if the d_tag ​​is the link, the link is "immutable" you change the description, which in this case is the content.

The d-tag will be the unique identifier

https://github.com/nostr-protocol/nips/pull/1849

But what about when an URL change, redirects slow down, might be getting outdated and not working anymore at all and so. Bookmark tools such as pinboard and raindrop can handle this well related to its history, interactions and so.

you have to make another pin

Yes, so you will loose history. Won't there be a nip solution to keep the benefits and history?

in the current form, the one we are using now

published_at tag as in the long form

the history with a simple tag r, or "r", "","timestamp"

That’s up to clients to figure that out, just like other bookmark tools do.

I mean how do we find this out with Yumyume if there is no unique identifier for this

The d tag is the unique identifier and cannot be updated. If a client will integrate an 'update bookmark' feature, that will mean a new event will be created with a new identifier. Then it's up to a client to do some matching for whatever reason.

Maybe two kind is the best solution

simple link 39700 (without changing the spec) like https://github.com/symfony/symfony/

category link 39701 like https://github.com/symfony/symfony/blob/7.2/src/Symfony/Component/HttpFoundation/Request.php#L98

Or maybe a smart relay which can figure that out and provide that (extra) metadata around those events

URLs break, that's what the internet does. There is no way around it. Even when URLs don't change the content inside them can change drastically.

One good solution is to keep a WARC copy of the referenced page stored somewhere, there is a DVM spec for that and I was working on such DVM but couldn't finish yet.

The generic solution is: make a new bookmark, keep the old one as a piece of history or delete it entirely.

If you're going to say a centralized system that is just a database can change the way it stores internal data and therefore it's better than fine, but Nostr is not such a thing and cannot ever have all these "features". Nostr is just a system that allows everybody to shout stuff so others can read, tradeoffs exist.