BIP 0015 development

[tgerring]

I'm interested in pursuing further development of BIP 0015, specifically in Namecoin.

I reached out to a few Bitcoin devs, but haven't received a response. What's the best course of action to champion this proposal and discuss implementations? I'd specifically like to involve wallet developers to gauge what's likely to actually make it into the clients.

[1714999346]

1714999346

[1714999346]

1714999346

[1714999346]

1714999346

[1714999346]

1714999346

[1714999346]

1714999346

[Mike Hearn]

It's largely superseded by the payment protocol, I think. The idea is to receive money someone sends you a file that contains all the information needed. That file could (with some extensions) result in a re-usable address book entry, although for now it doesn't due to the issues with address re-use.

[tgerring]

I'm looking to nudge a final specification into place to encourage other developers to leverage it, even if its not in mainline. Namecoin is the most promising IMO, but I'd like to propose a more standard key structure consisting of a hashed URI such that nearly anything can be referenced while discouraging blockchain scraping.

For example:

Code:

a-sha256/28439dee66dbbb66ace2de935adbce45c42625083b2f2cb3e7995a63f67fa5b0

Would represent the alias encoded as sha256. In this case, it happens to be mailto:foo@bar.net
It could also be used for SMS2BTC bridges, i.e. tel:+15558675309
Or custom namespaces, such as twitter:@ReallyPopularAccount

Some other things to consider:

• This would be a huge benefit to users, because they could register their own "green" address instead of relying on the wallet address of a third-party service to facilitate transactions, a la coinapult.
• Because Namecoin is distributed on a public blockchain in a manner similar to Bitcoin, entities can alias multiple pseudonyms whose contained information can never be suppressed by a central authority
• Name squatting is a possible issue. Can this be mitigated by signing the resulting hash with the corresponding private key for each public key they claim and storing that result as a measure of authenticity?

I'd love any feedback that moves us towards some consensus on this, so developers can start dreaming up interesting ways to apply it.

[WastedLTC]

I love the idea and feel NMC is the way to go.   You won't be able to keep others from adding records but you can add a hash that can be verified by a public key which will allow developers who implement the name look up within a wallet to only use values that can be validated with the shared key.  

If you setup a method to allow users to manage and add values to the NMC blockchain then sign the message. Follow that up with some acquiring trust in your process--  I don't see why this wouldn't catch on.  

Maybe create a workflow for each type of alias and a secure way for the user to validate.

example:  a domain btc alias can be validated by adding a file to the root of the website.  Your system finds validates the info and posts to the NMC block chain...   emails can be validated through confirmation clicks and so on.  

The fact that you could store a lookup for every currency, you can start on the alt chain and move your way up.  Sounds fun!

[domob]

I also think that Namecoin is a good way to store bitcoin addresses, at least for someone to publish his/her donation address or so.  See also https://bitmessage.org/forum/index.php/topic,2563.0.html, which does something similar for Bitmessage addresses.  We already thought about an experimental patch to a wallet (Electrum to be concrete) which does that - just replacing a namecoin name with the stored bitcoin address on request.  Is that what you were thinking about?

Note that this can even be extended to allow for non-public and non-reused addresses with a little more effort:  Instead of a static address, you can for instance also store the URI of a script on your server, which when queried returns a fresh address you can use.  This address can be signed by some "master key" stored also in namecoin to prove it is really owned by the person you try to send money to.

[tgerring]

We already thought about an experimental patch to a wallet (Electrum to be concrete) which does that - just replacing a namecoin name with the stored bitcoin address on request.  Is that what you were thinking about?

To some degree yes, except I'd prefer for the values to be stored in Namecoin in a standardized way. i.e. Not an arbitrary a/foobar or id/foobar, but instead a hash of the URI as proposed above. My thought being, the more clients that use a standard format, the better interoperability for everyone.