[]go1dfish in whatever

with that approach, only one key ever updates for any single id update

    []go1dfish in whatever

    and that key isn't 40k in size

      []go1dfish in whatever

      also by storing the sort values in the listing data, I avoid having to go out to redis for each item checked in a binary search to locate where to insert

        []go1dfish in whatever

        I only have to grab the listing itself, and the data necessary to calculate the sort value for the one item (like the timestamp or vote count or both)

          []go1dfish in whatever

          I'm thinking all together this will bring most living updates down to 20-30ms or less with minimal IO needed

            []FrontPageGuardian in whatever

            Oh and can you consider using sha1 on json instead of using object-hash as it would make thingId independent of js

              []go1dfish in whatever

              listing*

                []go1dfish in whatever

                I'd have to make some decisions about whitespace and stuff for that

                  [] in whatever

                  what is this thing

                    []go1dfish in whatever

                    it's not a bad idea though if you propose a better hash approach for things I'd definitely consider it and do a legacy thing in the schema like I do for the other old formats at least for a while

                      []go1dfish in whatever

                      This is an in development reddit alternative

                        [] in whatever

                        thats pretty cool

                          []go1dfish in whatever

                          built to be decentralized, with no single entity capable of unilateral censorship while empowering end users to have maximum control over their own experience

                            []go1dfish in whatever

                            Glad you think so :)

                              [] in whatever

                              how does the p2p thing work

                                [] in whatever

                                where is the info stored

                                  []go1dfish in whatever

                                  Notabug is built around this database standard: https://gun.eco/#step1

                                    []FrontPageGuardian in whatever

                                    Isn't just hashing json string of desired data enough?

                                      []go1dfish in whatever

                                      right now info is stored in superpeers like the notabug.io server

                                        []go1dfish in whatever

                                        as gun progresses, it will be possible to mirror and help contribute to network health via webrtc and leaving your browser open with indexeddb for storage if you choose to

                                          []go1dfish in whatever

                                          or a installed app that would run like a superpeer.

                                            []go1dfish in whatever

                                            In general it works around the idea of peers widely replicating data, but unlike blockchain there is no requirement that they replicate or keep all or any specific bits of data

                                              []go1dfish in whatever

                                              @FPG you would have to specify how to treat whitespace for that

                                                []go1dfish in whatever

                                                I know also gun is thinking about non-json transports at some point as wel

                                                  []go1dfish in whatever

                                                  so I don't know if I'd want to stick that close to the messaging format for defining the hash. currently it's more about the structure of the object in theory but yeah objet-hash could probably be better standardized or something