"human-readable" CID form should match spec

[olizilla]

...or spec should match what this site does.

https://github.com/multiformats/cid#human-readable-cids

# corresponding human readable CID
base58btc - cidv1 - raw - sha2-256-256-6e6ff7950a36187a801613426e858dce686cd7d7e3c0fc42ee0330072d245c95

https://cid.ipfs.io/#zb2rhe5P4gXftAwvA4eXQ5HJwsER2owDyS9sKaQRRVQPn93bA

Human readable CID
base58btc - cidv1 - raw - (sha2-256 : 256 : 6E6FF7950A36187A801613426E858DCE686CD7D7E3C0FC42EE0330072D245C95)

[lidel]

I'm 👍 for making things more clear, but sha2-256-256- is confusing: a reader does not know where hash name ends, and where length starts without a double take.

(a : b : c) removes ambiguity + makes it clear which segments are part of multihash, but perhaps there is better way of showing this? I think we should update specs and make them more clear, and them make cid.ipfs.io follow it.