Adam Langley <agl at linuxpower.org> writes: > Well, we have the metadata spec - and we want/need to encode some of > this into keys to avoid indirect lookups. So can we put keys with > metadata commands into Redirects etc? Messy. > > What do we want? In the end we want to return data and > true-metadata. Where true-metadata is what the rest of the world > thinks of metadata - e.g. Dublin Core stuff not Redirects and the > like. Since we've used metadata for pretty much everything under the > sun so far we need a new term: and I'm going to use "document meta > information" (DMI) for no very good reason except I need something > (and we get a new TLA).
The distinction that needs to be made, and which I'm not sure you're making here, is between "information about the key" and "information about the document". It's not at all messy to put information about the key into the key URI, and it's not messy to include a key URI in a Redirect. We already do both of those things. > In the real world > ----------------- > > Well, we need this in keys: > > freenet:SSK(aabbccddee,DBR(86500,site/index-)) > freenet:KSK(DBR(86500,mysite/index-)) Parentheses? In keys? Excuse me, I need to cry for a moment. > This is a long post for me. This metadata system at least looks > clean for the moment. It might need a Version header replacement. Sorry Adam, this metadata system is horrible. It needs a complete replacement. -S _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
