byjp, 4 months ago

@stevenaleach Exciting!
I think using CIDs as filenames might be confusing tho, as different CIDs can come from the same data (eg. you use CIDv0 but some future server sharing it uses CIDv1; or if someone switches from SHA256 to Blake3; or if that server uses different settings to yours).

The #InternetArchive has a non-functioning #Filecoin (storage) and #IPFS (transfer) option for retrieval they’ll hopefully be fixing, perhaps uploading with human filenames would be better?

travisfw, 4 months ago (edited 4 months ago)

@byjp @stevenaleach
Highly relevant: RFC 6920: Naming things with hashes
https://datatracker.ietf.org/doc/rfc6920/?include_text=1
IE if you don't mind creating a directory structure, you could include in a root folder "ni", directories named for the hash algorithm. Then the files in each directory will be named, obviously, the hash the algorithm produces. A few easy examples in the "Hello World" in section 8.1
https://datatracker.ietf.org/doc/html/rfc6920#section-8.1