Could we further modularize the protocol?

[danieldaquino]

I like the concepts used in the Blossom protocol, and it is similar to a NIP I was working on (https://github.com/nostr-protocol/nips/pull/1098/files)

I was considering turning that NIP draft into its own protocol (Blobstr), but I believe it might be better to implement some of those ideas into the Blossom protocol instead to avoid duplicating efforts.

However, the concern I have about the Blossom protocol is that BUD-01 is a bit monolithic and constraints the possible use cases. For example, it prevents the following:

• Read-only Blossom servers (e.g. Say someone just wants to host a curated set of files, or share files from a local directory, as some form of archive)
• NIP-98 authorization (e.g. If users want to purposefully limit the upload/download to specific servers for better privacy, for example in the case of operating within a local network)
• Likely other use cases

In general, content-addressable storage can serve a big variety of use cases, so I believe it might be a good idea to reduce BUD-01 into something smaller and add-on the other features/requirements into BUD-02, BUD-03, etc, to serve specific use-cases, that way the protocol can serve more use-cases. If BUD-01 is too restrictive and mandatory, it cannot be used for a broad set of use cases and will encourage those to use different protocols.

Is this something you'd be interested in? I could come up with some drafts if interested.

[hzrd149]

So maybe something like #15

[danieldaquino]

Thank you @hzrd149 for putting that together! Yes, that's what I had in mind.

I still need to wrap my mind around the GET authentication/authorization mechanism, but I can open a separate ticket if I have concerns about that