I'm thinking about building a distributed GLBSE
(if enough developers join in - I don't have the time to do this alone). I think an implementation could be built on top of Bitcoin & Namecoin, instead of a separate blockchain.
So, imagine this: To register a new asset, you buy a long Namecoin address that is the signature on a contract (published into the bitcoin blockchain itself as a text message). You might want to have 1,000,000 different stocks of this asset, that will each require a Namecoin address in order to be uniquely transferable (let's not worry about stock splits for the moment).
Currently, this might be very expensive, since each Namecoin registration costs a given amount of money, and 1M of them will cost a lot.
Is there a feasible way to change the Namecoin protocol such that long addresses/names will cost less to register?
This seems fair to me, since there a lot more long addresses, so the "inherent value" of a long address is smaller than a short one (compare the "abc.com" domain to "21z980347sd9fhase43wsadf.com")
Would anyone even consider making such a change in Namecoin? Would this cause an influx of spam in the Namecoin blockchain? There would be a minimal fee for an address, of course, it would just be lower than the current fee (which is what btw?)
The reason I think this should be implemented on top of current chains is that they already have a large hash rate, and any new chain we start will be prone to 51% attacks until merged mining is introduced ... which won't be for some time. If this can be reasonably implemented on top of existing chains, we benefit from all the existing hash power from day 1.