Hi all,
The canonical unit of data interchange in our P2P client is a resource,
which basically conforms to the RDF data model. So each resource has a
unique ID, which is a URI currently composed of a unique "peer ID"
assigned by our registration server and a unique ID for that peer (using
a counter) assigned by the peer itself. We use the "allpeers" scheme to
indicate that the URI is using our proprietary format. For example, if
my peer ID is 1234, a given resource that I create might have the ID
allpeers:1234:5678.
We would like to support other unique IDs, particularly for file
metadata using the file hash. So we're considering supporting IDs like
sha1:XYZ and md5:XYZ. However, we noticed that file-sharing clients tend
to use the "urn" scheme if they are going this route. For example,
Azureus magnet links look something like urn:sha1:XYZ. I guess this is
because in some contexts a "known" scheme is needed, but I'm curious to
know the details. Does anyone know why people prefer to use URNs rather
than URIs with the hash type used as the scheme directly? Or is there
some generic processing of hashes possible that might lead us to prefer
a format like hash:sha1:XYZ and hash:md5:XYZ?
Thanks for any input,
Matt
_______________________________________________
p2p-hackers mailing list
[email protected]
http://lists.zooko.com/mailman/listinfo/p2p-hackers
