search

ceramicnetwork / CIPs

The Ceramic Improvement Proposal repository
https://cips.ceramic.network/
MIT License
82 stars 22 forks source link

Discussion: DocID encoding #59

Open oed opened 4 years ago

oed commented 4 years ago

Discussion for CIP-59.

michaelsena commented 4 years ago

@oed This looks great, and very useful especially as we have more than one network. On that note, is namespace a standard in multiformats, or something we created. If it's something we created, then perhaps networkID is more clear? Unless you see a use case for this where the namespace will not equal a networkID.

oed commented 4 years ago

@michaelsena Its different from multiformats. I think namespace is more clear since networkID is usually (in ethereum) used to refer to the peer network itself. Where as this is actually more of a namespace for documents.

ukstv commented 4 years ago

To clarify, which base32 is used here? Also, padding?

oed commented 4 years ago

@ukstv See here: https://github.com/multiformats/multibase (updated spec with link)

OR13 commented 4 years ago

can i get a link to the format of the complete bytes for the resource identified by genesis-cid-bytes? (I want to see what genesis data is)

oed commented 4 years ago

You can have a look in the Ceramic specs @OR13 Note however that different doctypes might encode this using various ipld codecs. For example, tile uses dag-jose, while 3id and caip10-link just uses dag-cbor.

In theory a doctype could support a genesis format of any shape, as long as ceramic has a way to interpret what doctype is being used.

oed commented 4 years ago

Updated the specs and created a PR: https://github.com/ceramicnetwork/CIP/pull/61

oed commented 4 years ago

@zachferland can you please make a PR to update the DocID spec based on the changes that ended up in the implementation?