BIP 0020

From Bitcoin Wiki
Revision as of 12:44, 10 January 2011 by Marcusaurelius (talk | contribs) (use-cases)
Jump to: navigation, search

This is about creating a URI scheme for bitcoin. Previous discussion in the forum x-btc specification at

RFC 3986

the following is taken from wikipedia

Internet standard STD 66 (also RFC 3986) defines the generic syntax to be used in all URI schemes. Every URI is defined as consisting of four parts, as follows:

<scheme name> : <hierarchical part> [ ? <query> ] [ # <fragment> ]

The scheme name consists of a letter followed by any combination of letters, digits, and the plus ("+"), period ("."), or hyphen ("-") characters; and is terminated by a colon (":").

The hierarchical part of the URI is intended to hold identification information hierarchical in nature. Usually this part begins with a double forward slash ("//"), followed by an authority part and an optional path.

  • The authority part holds an optional user information part terminated with "@" (e.g. username:password@), a hostname (i.e. domain name or IP address), and an optional port number preceded by a colon ":".
  • The path part is a sequence of segments (conceptually similar to directories, though not necessarily representing them) separated by a forward slash ("/"). Each segment can contain parameters separated from it using a semicolon (";"), though this is rarely used in practice.

The query is an optional part separated with a question mark, which contains additional identification information which is not hierarchical in nature. The query string syntax is not generically defined, but is commonly organized as a sequence of <key>=<value> pairs separated by a semicolon[1][2][3] or separated by an ampersand, for example:

Semicolon: key1=value1;key2=value2;key3=value3
Ampersand: key1=value1&key2=value2&key3=value3

The fragment is an optional part separated from the front parts by a hash ("#"). It holds additional identifying information that provides direction to a secondary resource, e.g. a section heading in an article identified by the remainder of the URI. When the primary resource is an HTML document, the fragment is often an id attribute of a specific element and web browsers will make sure this element is visible.


I propose a scheme like this:

() means optional, $* are placeholders



label: Label for that address (e.g. name of receiver)
address: bitcoin address
query: pairs of key=value seperated by &
message: optional message that is shown to the user after scanning the QR code

Query keys

amount: amount of BTC


Just the address:


Address with name:


Request to send 20.30 BTC to me:


Request to send 50 BTC with message:


Characters must be URI encoded.

tcatm, modified by LukeJr

I propose a scheme like this:

[] means optional, <> are placeholders



  • label: Label for that address (e.g. name of receiver)
  • address: bitcoin address
  • message: optional message that is shown to the user after scanning the QR code
  • size: amount of $unit
  • unit: BTC, [ᵗˢᵐᵇ]TBC[ᵗˢᵐᵇ], or null/omitted (which refers to the base unit)


Just the address:


Address with name:BTC


Request to send 20.30 BTC to tcatm:


Request to send 400 TBC to Luke:


Request to send 5 uBTC:


Request to send 50 BTC with message:


Characters must be URI encoded properly.

BNF syntax

bitcoinurn      = "bitcoin:" bitcoinaddress [ ";version=" bitcoinversion ] [ "?" bitcoinparams ]
bitcoinaddress  = FIXME :)
bitcoinversion  = "1.0"
bitcoinparams   = *bitcoinparam
bitcoinparam    = amountparam | labelparam | messageparam
amountparam     = "amount=" amount
amount          = amountdecimal | amounthex | amounttonal
amountdecimal   = digits btcunits
amounthex       = "x" hexdigits bitcoinunit
amounttonal     = tonaldigits tbcunits
bitcoinunit     = btcunits | tbcunits
btcunits        = "BTC" | "BTCents" | "uBTC" | "uBTCents"
tbcunits        = tbcunitsunicode | tbcunitsascii | tbcunitsfull
tbcunitsunicode = "ᵇTBC" | "ᵐTBC" | "ˢTBC" | "ᵗTBC" | "TBC" | "TBCᵗ" | "TBCˢ" | "TBCᵐ" | "TBCᵇ"
tbcunitsascii   = "bTBC" | "mTBC" | "sTBC" | "tTBC" | "TBC" | "TBCt" | "TBCs" | "TBCm" | "TBCb"
tbcunitsfull    = "tran-TBC" | "song-TBC" | "tam-TBC" | "bong-TBC" | "mill-TBC" | "san-TBC" | "ton-TBC" | "TBC" | "TBC-ton" | "TBC-san" | "TBC-mill" | "TBC-bong"
labelparam      = "label=" *uchar
messageparam    = "label=" *uchar


payment identifiers, not person identifiers

in my opinion, the most basic idea of the URI scheme (as this is a currency) is to facilitate payment. so the URIs should represent first and foremost payments. if it represents something else, this needs to be specified. Thus bitcoin:13guMzcGPvdD3qjQvCoNc1w5XAgJ638KaQ represent a payment to me using my bitcoin adress, not my bitcoin adress itself. So after parsing the URI (via link/qr/whatever) the application should open a transaction window with the adress filled in. you then need to add an amount and confirm the payment. If your application is smart, it will also have a button "just store the adress". But the point i am trying to make, is that the default use of the URI should be for payment, nor for exchanging adresses.


imported from the forum: I like the simplicity of bitcoin:xxxxxxxxxxxxx plus very much approve of its accessibility. Should someone from the outside happen to see such a uri, the protocol name already gives a description. A quick google search should then do the rest. x-btc sounds much more cryptic, the chance that s/1 gooogles that out of curiosity are much slimmer. Also, very likely, what s/he will find are mostly technical specifications. Not a good introduction to bitcoin.

for the same reason i am for using '&' as a delimiter for key-value-pairs. people know it from urls. make it easy for people to understand what is going on.

keep it simple

don't explicitly write down information that can be inferred. don't mark the address as an address. if there is no address, this does lose much of its utility. we could, however, specify 'address' as a reserved word, so that bitcoin:address?amount=50 would initiate a transaction with the amount filled in, but with a blank address. I am not convinced that there is a use case, though.


before the URI scheme is finalised one should think long and hard about use-cases. in what circumstances will who use this for what?

  • an online shop has a 'buy this'-link, which uses the URI scheme.
    • PROBLEM: click on the link opens the application. how does the merchant notice this.
      • POSSIBLE SOLUTION: javascript can detect the click.
      • POSSIBLE SOLUTION: the checkout site checks the httprequest
    • PROBLEM: the time problem (~10minutes) is very apparent here. nobody wants to wait 10minutes for the transaction to be confirmed.
  • a person only has an online client, no actual application
    • PROBLEM: how to redirect the URI, so that the online client gets a notice?


we want URIs generated in 2011 to still work in 2036. think about extensibility. of course we can make only educated guesses (and nothing more!) about the future, but don't act as if there is none. this should be the best we can do, but it should not be seen as forever set in stone. make it possible for later generations to improve our work, to mend our errors, without breaking the URIs created now. Version incompatibility is the easiest thing to drive users crazy: "i now upgraded to this shiny new version. what? it doesn't support the old format? AAAAAAARRRGH!"


  1. RFC 1866 section 8.2.1 : by Tim Berners-Lee in 1995 encourages CGI authors to support ';' in addition to '&'.
  2. HTML 4.01 Specification: Implementation, and Design Notes: "CGI implementors support the use of ";" in place of "&" to save authors the trouble of escaping "&" characters in this manner."
  3. Hypertext Markup Language - 2.0 "CGI implementors are encouraged to support the use of ';' in place of '&' "