fs-like namespace design

similar to ndn i'd like to operate in terms of a unix-like hierarchical namespace like

/ssb/ed25519/c6eab901347a84336debfd2ad5107e2f57f4772795935906d49017bb88e8fc65
/ssb/blob/sha256/77c8b411f04f12e775ed36dd386629927c12ba10b4da48fdb7cfd73329c858d0
/ap/social.sunshinegardens.org/notice/9ieXEjBcRWduUNOmO0
/ap/dickkickextremist.xyz/objects/c036812a-0270-4f52-b419-864df4c58681
/web/https/xj9.io/tags/technology/

• https://github.com/libfuse/python-fuse

with 9p maybe? does that make sense as a generalized ipc interface?

perhaps the interface we care about is symbolic. w3c uris are pretty complicated and it would be nice to have a universal way to identify resources that is simpler. 9p and NDN are concrete examples of fs-like network systems, but the syntax itself is pretty nice.

/ssb/feed/ed25519/<feed id>
/web/xj9.io/cvcvcv
/tracker/pki/<id>

its pretty easy to map from w3c uris too. there's a low limit to name complexity, but that's probably a good thing. there are some questions about how name registration happens, but its probably good enough for people to define and trust namespaces on their own.

i think 9p is a good enough fit to start pursuing this idea. one question that comes to mind is, how well can 9p be made to work over LoRa or other slow wireless links? perhaps an additional translation layer is needed for that. moving away from 9p in the future is fine as well, but its a good starting point for thinking about filesystem-like representations of p2p networks.

i don't necessarily want to build another plan 9 system, but there are good ideas in there that i'd like to incorporate into this project.

the next actionable item is porting py9p to micropython/pycopy so it can be used with gopheric to fetch information provided by another process that'll act as a stand-in for sora for testing. a good place to start could be serving the config that gopheric is currently reading from ${SORA_HOME}/*.json.