[bitcoin-dev] Human readable checksum (verification code) to avoid errors on BTC public addresses

ts ts at cronosurf.com
Thu Aug 19 17:02:38 UTC 2021

Hello ZmnSCPxj,

ZmnSCPxj wrote on 8/16/21 5:34 AM:
> Good morning TS,
>> Entering a BTC address for a transaction can pose a risk of error (human or technical). While
>> there is a checksum integrated in BTC addresses already, this is used only at a technical
>> level and does not avoid entering a valid but otherwise wrong address. Moreover, it does not
>> improve the overall user experience.
>> In case this hasn't been discussed before, I propose to implement a 3 or 4 digit code (lets
>> call it 4DC for this text), generated as checksum from the address. This 4DC should be shown
>> in all wallets next to the receiving address. When entering a new address to send BTC, the
>> sending wallet should also show the 4DC next to the entered address. This way, the sending
>> person can easily verify that the resulting 4DC matches the one from the receiving address.
>> This would mean that a receiver would not only send his public address to the sender, but also
>> the 4DC. A minor disadvantage since a) it is not mandatory and b) it is very easy to do.
>> However, it would greatly reduce the probability of performing transactions to a wrong address.
>> Technically, this is very easy to implement. The only effort needed is agreeing on a checksum
>> standard to generate the code. Once the standard is established, all wallet and exchange
>> developers can start implementing this.

Thanks for your comments.

> I think the "only" effort here is going to be the main bulk of the effort, and it will still take years of agreement (or sipa doing it, because every review is "either sipa made it, or we have to check *everything* in detail for several months to make sure it is correct").

I understand. If sipa could do it that would greatly simplify the process. Once an algorithm 
for the generation of the code exists, it just needs to be communicated to wallet developers 
and let it grow organically. No need of extensive testing, since it is only a very simple 

> In any case --- the last 5 characters of a bech32 string are already a human-readable 5-digit code, with fairly good properties, why is it not usable for this case?

Well, because
a) most people don't know that
b) it is specific to bech32
c) it is not easily readable being the last digits of a long address (although this could be 
fixed by the wallet by showing those digits bigger or separately)
d) and most importantly: as mentioned in above my proposal, it only proves that an address is 
valid, but not necessarily the correct one (perhaps the user copied the wrong address, there 
was an old address in the clipboard, etc.)

> On the other side of the coin, if you say "the existing bech32 checksum is automatically checked by the software", why is forcing something to be manually checked by a human better than leaving the checking to software?

Not better, it should be on top. And not forced, but just as an optional check for the user. 
The SW can (and should) only check that the address is valid (the SW doesn't know the user's 
intent). Only the human can "double-check" an easy-to-read-code to quickly know that he is 
doing the right thing. (Entering a valid but wrong address is even worse than entering an 
invalid one, since the latter will be stopped by the wallet. But the former most likely 
results in loosing the funds.)

Note: The code should never be entered manually or even copied together with the address in 
one string. From the SW point of view, the code is an output only, never an input. It is 
merely a visual verification for the user.

Example of use: person A calls via phone person B and says: "Send me 0.1 BTC to my address I 
just sent you via whatsapp. When entering the address, make sure that you get the verification 
code 4385."


> Regards,
> ZmnSCPxj

More information about the bitcoin-dev mailing list