so i've started writing a little pastebin style clone today, with an interface over HTTP and storage backed by IPFS: https://github.com/grufwub/gibon 

already got a very simple prototype working :O

until last night i really had no idea how it all worked, now i have a basic idea of IPFS and libp2p and it's very cool!

I'm doing this directly by block storage so i'm not totally sure of the implications of this. my first thoughts right now are to try uploading something > 256KB see if it automatically chunks it as i would expect

if not then i may need to use the filesystem api

i just figured if i was dealing with byte slices anyway why not write directly to block storage

okay yeah i've just made a simple HTTP gateway so you can access any number of /ipfs, /ipld etc

so have put a 1mb data restriction in place, and only support putting of linked data directly

*only support POSTing and GETting

though if i ever make an actual web interface and support syntax highlighting, switching to the filesystem API would be easier since i can use specified filename as a means of picking syntax highlighting

if someone actually knows IPFS, please feel free to step in here and point out anywhere in this thread (or in my code) where I'm being a dumbass :)

i used the various IPFS sub-repo godocs, and inspiration from here for node setup: https://github.com/ipfs/go-ipfs/blob/master/docs/examples/go-ipfs-as-a-library/main.go

it was really a lot easier than i thought