As I said Mar 15, 2001 and other posts,  there seem to be many
types of decentralization, these are independent dimensions

- the data,
- pointers and indexes to the data,
- other metadata i.e. useful metadata like document schemas,
- the encryption keys and certificates,
- routes taken by data content across networks (P2P variations),
- routes taken by the pointing, indexing, metadata information,
- incentive topology ($ Corp, $hierarchic/MLM, $peer, mojo, volunteer...)
- capital structure (pooled/central, decentralized, or hierarch. steps)
- political and social rulemaking topology /hierarchy,
- actual technical control (actual sovereignty)
- content creation (added by Mitra)

It's amusing to consider different combinations.  e.g what if the
data stays central but the indexes are distributed?   Or what if
the economic incentives are only positive for centralization? how can
we make them positive for decentralization in a way that grows a
physical network that maximizes effective sovereignty at the
nodes etc. (if that is one's goal)

The number of permutations are fairly numerous considering that
decentralization has so many additional colors and flavors, and can
exist in separate universes or contexts like wireless networks, dialup
BBS universes like FIDONET, as well as the cisco/telco IPV4 internet
or the PSTN.

Too often on this list, messages assume a particular context.  We need
an organizing principle.  That principle is, what result you're trying
to accomplish.  Then, the discussion is more productive.