3 ms·

I wrote the software used at https://loom.cc https://loom.cc. That is a site which allows anyone to create and issue new types of digital assets. Typically th

by fexl 16y ago

I wrote the software used at https://loom.cc https://loom.cc. That is a site which allows anyone to create and issue new types of digital assets. Typically the issuer will write a contract pledging to redeem the digital asset for a specific quantity of physical assets or services. Owners of the digital asset may spend it freely on the server, similar to how they spend cash in the physical world. One thing I like about Loom is that an asset type has a single issuer. This means that an issuer can store some physical assets in a vault, create a digital asset type redeemable for those physical assets, and issue the digital asset type in a strictly limited quantity. Then there's no question what a digital asset type is good for, because even if no exchangers or merchants accept the asset type, at least the issuer is contractually obligated to deliver the assets in storage or the promised services. I wish Mr. Nakamoto (the author of Bitcoin) could devise a way for a digital currency to have a single issuer, but still trade in the wild like Bitcoin does now without a central server. However, that may be a logical contradiction and thus fanciful thinking. So the best I've been able to come up with is Loom, which does indeed depend on a central server. But I would like to see many Loom servers with trading networks between them. That might be a way to realize the benefits of central servers while avoiding much of their risk. (invite 42215c198e22c724d7cfd7887bda2a57) P.S. Various source code references here: https://loom.cc/view/eae93b14a717a7f43e6354bade2a199ca2e3bef4971eddc5dba1636f026c94e7 https://loom.cc/view/eae93b14a717a7f43e6354bade2a199ca2e3bef...

eru 16y ago
I heard of cryptographic protocols that can achieve what you ask for here. (I just have to dig up where I read about it.)
- fexl 16y ago
  I've read up a bit on blinded signatures, Brands, etc. I have two problems there. First, I would need a reference standard implementation of blinded signatures. I know if I tried to cobble something together myself, some crypto-guru would laugh and point out that my prime factors weren't super-prime in a Zeta field of modulus three or some such thing. Second, basic blinding still does require a central server. I believe the Brands protocol allows blinded tokens to circulate without contact with the server at every stage, and anyone who attempts a double-spend can be identified cryptographically. But it's complicated and ultimately relies on the technique of going after the bad guy with a baseball bat. So if I went with blinded tokens on a central server, I'd be right back at the central server problem. Blinding provides un-traceability at the cryptographic protocol layer. With Loom, I figured that anyone that concerned with un-traceability could just use a VPN or Tor.
  - eru 16y ago
    Thanks for the write up!
  - derefr 16y ago
    What about a peer-to-peer solution? Every participant runs a brand server; to "spend" a token, you contact any random server and tell it that you're doing so. That notice then gets spread to every other server; a server who has received the notice won't allow the same token to be spent again, and a server that isn't "caught up" to the mesh won't be allowed to be authoritative (be a source for spend announcements.) It would basically be like Bittorrent—except the pieces are created after the torrent is, inserted into the "file" (keyspace) one at a time, and the hash mutates with every spend.

by fexl 16y ago

eru 16y ago

I heard of cryptographic protocols that can achieve what you ask for here. (I just have to dig up where I read about it.)

fexl 16y ago
I've read up a bit on blinded signatures, Brands, etc. I have two problems there. First, I would need a reference standard implementation of blinded signatures. I know if I tried to cobble something together myself, some crypto-guru would laugh and point out that my prime factors weren't super-prime in a Zeta field of modulus three or some such thing. Second, basic blinding still does require a central server. I believe the Brands protocol allows blinded tokens to circulate without contact with the server at every stage, and anyone who attempts a double-spend can be identified cryptographically. But it's complicated and ultimately relies on the technique of going after the bad guy with a baseball bat. So if I went with blinded tokens on a central server, I'd be right back at the central server problem. Blinding provides un-traceability at the cryptographic protocol layer. With Loom, I figured that anyone that concerned with un-traceability could just use a VPN or Tor.
- eru 16y ago
  Thanks for the write up!
- derefr 16y ago
  What about a peer-to-peer solution? Every participant runs a brand server; to "spend" a token, you contact any random server and tell it that you're doing so. That notice then gets spread to every other server; a server who has received the notice won't allow the same token to be spent again, and a server that isn't "caught up" to the mesh won't be allowed to be authoritative (be a source for spend announcements.) It would basically be like Bittorrent—except the pieces are created after the torrent is, inserted into the "file" (keyspace) one at a time, and the hash mutates with every spend.