@kirby @FrailLeaf @Lyx @Owl @Owl @cassidyclown @drewdevault @kent @pernia Usually, you wanna store what you got from the other server verbatim. Pleroma just uses regular indexes but it uses a locally assigned ID (flake ID) in addition to the AP ID (the actor URL). I don't know exactly how Mastodon does it, but I know how Rails "programmers" do it and they probably almost exclusively use an in-DB sequential ID and have a pile of methods with names like "friendly_url" (and it generates the path component of the URL rather than an actual URL). Revolver just uses an index (named the "ID Index Orthogonality Table", because only a fool writes his own K/V data store), and it has some block IDs associated with it, and for AP entities, things like the timestamp of the last attempt to fetch, etc.