> Owning your own domain name is probably the only practical thing to do long > term if you want to control your data. Actually I do that already, but I do not expect all potential users want to buy a domain. So my approach was dreaming of somehow using public keys as identifier which could do the work of a domain in foaf. And from foaf one could relate oneself to people by using their keys :)
I mostly thought that it should be possible with a system similar to BitTorrent. Provide an entry point once and go from there using whichever seemingly magical technique. (I always thought that DHTs do this work, is this right?) > You can of course use services, and I think that will not just be tied > to Facebook or LinkedIn. Universities will certainly want to give their > students and employees WebIds, which people will be able to use as their > academic persona. Companies will want to give their employees WebIds to > communicate with their business partners. These WebIds may not necessarily > be tied to real people but could also be tied to roles in the company. But services tend to get evil at some point of time (Seen that 'Ning closing free ...' thread on the list? I think it was on social-discuss). I think my basic problem is that of URIs and that one has to have some ressource availlable all the time. But firstly I do not have an all-time connection, and secondly I would like to see something similar to a data cloud, but free and so on. I mean that my data gets mirrored by viewers when they first view it and that I could request someone else to mirror mine and as the network would grow, my data would finally be availlable in different places but not just from me. Unfortunately it is a quit complex theory as this would need a graph versioning system, proper encryption of parts of the data and even some method to simply drop outdated data. > HTTP is a p2p service. Just think of what would happen if you put a web server > on your computer. Opera I think if started playing with that idea, by putting > a web > server in the browser.... (note to self: Need to look at that in more detail) > > > Nonetheless, I think that it is important to > > support both, even if the latter would require a server and a domain. > > So you are ok with a domain, and a tcp connection! Welcome to HTTP :-) But not entirely happy. Because to me the p2p means that there is no client-server relationship. All nodes are equal and can do both parts. Possibly I have simply to wait some time for smart people designing dynamic p2p URIs that resolve to the most appropriate currently online source. But in fact, I can not wait...
