I've finally taken the time to read through the discussion. Is there an "official" IRC channel or something? If not, want to join #bitstorage at freenode?

Some points I want to add to the discussion:

- When I began thinking about this concept, I imagined a dropbox-like system, in which you can upload any kind of files. Later I decided it could be better to limit the files to a certain size, or forget about the notion of files altogether (at least at the protocol level). That's why I had named the system "bucketchain", as a system which allows distributed storage of 1MB-sized "buckets" of data. (I like bitstorage better, but we can still use "bucket" to refer to the unit of storage).

- Regarding the impossibility of having proof of redundancy (because a node can store a bucket 1 time and expose it as many identities), I read an idea that could be applied here. The client can allow the user to upload a file with n-redundancy by encrypting the file n times with different private keys. For the network that'd be n different files, so they can't cheat. The problem with that is that it can be stored n times in the same server, which removes part of the redundancy benefits. In any case, I don't think the protocol should contemplate redundancy. An consequently...

- We should try to reduce the protocol to the bare minimum, and then build over that. I think that'd be one in which you can create write transactions for 1MB buckets, and read transactions which are redeemed by peers who can prove to have that bucket. Actual bucket transfer can be left for a side-channel maybe?

- Is it too early to create an open-source project and github repo for this? I think Sarchar should start the thing, but I can do it too. I suggest python as the language. It makes no sense IMO to create a C++ project today. We'll get lots more collaborators and readable code using a more modern language.


Sorry for the long post!
Manuel