|
Title: Bitcoin DNS addresses Post by: wessorh on July 25, 2011, 11:56:13 PM Just got out of newby jail...
I'm working on a set of patches centered around leveraging the DNS to facilitate looking up a bitcoin address for sendtoaddress. Currently the server expects a bitcoin address like 1GCVXLfF1TcpnnDLJRHk845NZhuJWQTnUD These addresses are too hard for my friends to remember and I'd like to make it easier to use an address that looks much more like an eamil address. I have a patch to sendtoaddress so that it can take something like rick@wesson.us which is not an email address. In the DNS under wesson.us I've included a TXT record which can be resolved by prepending _btc and replacing the at (@) with a dot (.) so rick@wesson.us becomes _btc.rick.wesson.us which can be resolved to the following text... Code:
I'm working on two efforts, one to patch sendtoaddress to preform the lookup if the address has an at-sign (@) and to encourage wallet hosting folks to also put their userids in the DNS as a text record. The second effort is to write a internet-draft that explains how to use the TXT record for bitcoin wallets and describe a potential new resource record called the Wallet Address or WA I understand this effort will potentially erode some annomity for those that use it, and that what I'd like comments on. What is the social contract for ease of use -vs- security by obscurity. I'm also generally interested in comments on getting patches into the main line to support bitcoin adresses in the DNS. -rick Title: Re: Bitcoin DNS addresses Post by: jackjack on July 26, 2011, 12:21:01 AM Monitoring...
Also, if anyone doesn't want to appear in those DNS, nobody can force them so I don't see any problem about anonymity Title: Re: Bitcoin DNS addresses Post by: Serith on July 26, 2011, 01:08:50 AM Just got out of newby jail... I'm working on a set of patches centered around leveraging the DNS to facilitate looking up a bitcoin address for sendtoaddress. Currently the server expects a bitcoin address like 1GCVXLfF1TcpnnDLJRHk845NZhuJWQTnUD These addresses are too hard for my friends to remember and I'd like to make it easier to use an address that looks much more like an eamil address. I have a patch to sendtoaddress so that it can take something like rick@wesson.us which is not an email address. In the DNS under wesson.us I've included a TXT record which can be resolved by prepending _btc and replacing the at (@) with a dot (.) so rick@wesson.us becomes _btc.rick.wesson.us which can be resolved to the following text... Code:
I'm working on two efforts, one to patch sendtoaddress to preform the lookup if the address has an at-sign (@) and to encourage wallet hosting folks to also put their userids in the DNS as a text record. The second effort is to write a internet-draft that explains how to use the TXT record for bitcoin wallets and describe a potential new resource record called the Wallet Address or WA I understand this effort will potentially erode some annomity for those that use it, and that what I'd like comments on. What is the social contract for ease of use -vs- security by obscurity. I'm also generally interested in comments on getting patches into the main line to support bitcoin adresses in the DNS. -rick I think that won't work, because recommended practice for a merchant is to use new address for every incoming transaction to verify that the payment was received from a customer https://en.bitcoin.it/wiki/Merchant_Howto#Common_Errors (https://en.bitcoin.it/wiki/Merchant_Howto#Common_Errors) . Title: Re: Bitcoin DNS addresses Post by: wessorh on July 26, 2011, 02:46:24 AM There are several options for merchants, my first goal was to address my friends ;-)
For the merchant simply including an identifier to map your transaction back to. One might include this identifier in the "message" portion of the transaction. The identifier could be anything, your street address for pizza delivery, your email (pay-from) address which could be resolved from the DNS (again) or a simple pin. If the transaction fees are less than current electronic transactions and easier to use than cash I'd say both the merchant and the customer win. maybe I'm wrong with the message space being used to identify the sender, but both the sender and merchant would benefit by ease of use. If the merchant can look up the address to expect a transaction from then haven't we addressed your issue. (M)erchant (C)ustomer C: I want a pizza M: takes order, informs C to pay X to merchant.example.com C: ok, its coming from customer.example.com M: have a drink, its on the house this might a day to clear <time passes> M: sees txn from <address> verifies address is associated with customer.example.com C: takes pizza In this case the "message" field of the transaction wasn't populated, but it could have helped. Looking forward to learning if the message field can be leveraged for the above. |