>>21I'm just measuring the benefits and draw backs of using something like HTSQL to just proxy sql commands over a network, or providing a more special purpose protocol that could be safer for keeping synchronization but also more restrictive and reinvented. The concept of executing SQL remotely gives the most flexibility, but may push too many synchronization details into the application. But maybe these are necessary. I was hoping to abstract that sort of stuff so the applications would be easier to write.
For the anonymity stuff I'm just relying on Tor for now. Since tor can act pretty transparently, I'm not putting any details like that into it and just assuming some kind of ip address or domain name is resolved and a TCP connection is created.
Also I have distbb in mind for the first application of this, but I want to generalize it other sorts of websites that are 2.0 and typically centralized.
>>22Yeah, a rouge client could probably mess stuff up. The read only SQL commands would be good though. If the client wants some specific information about a peer's database, if the protocol can't specify a way to get it, the client will get the information by retrieving more information than necessary, which will put more load on network.