We should discuss if we need Paymail

I am not a Bitcoin expert. Like all my articles this is a subjective take and an exercise at thinking through an idea well enough to put it in written form. It is also entirely possible that it is wrong. Do not ask use this article to justify questioning Paymail in any wallet — they are smart enough to come to their own conclusions!

Please do not respond to this on Medium as I prefer not to use their comment system, discuss this with me on the Metanet.ICU Slack or the BSV Discord.

When Paymail was broached at the first wallet workshop, it seemed like a good idea. I went away and wrote a Paymail hosting service for ElectrumSV’s users, implemented a contact system for Paymail addresses, the required user interfaces and more. ElectrumSV was probably going to be the first wallet to support Paymail, all that was required was for the patents to be granted and Paymail to be announced.

You might be aware that near on four years later ElectrumSV somehow still has no support for Paymail. So, why is this and what do I think about Paymail? I’ll try and describe my thoughts on Paymail below, but keep in mind I am not a Bitcoin expert, or an expert in anything for that matter. Maybe I am completely wrong!

Some of the problems

Meet the new boss, same as the old boss

What Paymail at the surface level seems to provide, is a way for an identity to be hosted on your behalf. In theory anyone can run their own servers and host their own Paymail addresses, but the modern services 99% of people rely on show that this is both prohibitively complicated and not a viable proposition. The core businesses that provide the services (Google, Facebook, etc) sometimes end up being the only options for “identity”, although most but not all web sites still support email-based sign-up which is pretty much a different form of the same thing. After all your email address is going to be at one of the same places even if you log in with email rather than linking your account!

Paymail is a way for a business to say, you are my customer and this is the way that others interact with you on my business platform. It is unclear what value it could add in the peer to peer world, where the user controls their own data, keys and coins. It is not an identity, but an account at that business and one that is is advertised by showing use of unfiltered unsolicited payments in an unsecured way.

Paymail is … an account at that business and one that is is advertised by showing use of unfiltered unsolicited payments in an unsecured way.

Liability issues

Paymail is more akin to a postal service than a bank account. You are not directly paying the person the Paymail address represents, you are paying through the service that is hosting that address and maybe paying that person. Or worse you are paying a hacker who has been waiting for large payments to come through — keep in mind that even Apple and Microsoft which are huge multi-national corporations get hacked. The hacker can sit there waiting and then pick off selected payments and the hosting service or wallet might take a while to work out what is going on. They’d probably be more likely to blame the sender not sending the payment than detect the hacker, until a large amount of damage is done.

A wallet that supports payment to Paymail addresses should inform the user that a payment to “bob@bobberson.com” is not guaranteed to reach “Bob” and they send at their own risk. It’s what I was going to do. If they do not warn the user then either they are choosing not to do so to avoid the unpalatable sheen it gives Paymail, or they are not technically skilled enough to realise this is the case. Whichever the reason is, that wallet might find themselves liable by the paying user whose payment did not go where the wallet said it would.

If you are a Paymail hosting business and you host “bob@bobberson.com” and you are hacked and a payment goes missing, your best bet is to deny you were hacked or that the payment ever reached your service. In fact, you might even think that that is what happened. However, if it is proven because the paying wallet keeps records and receipts of some sufficient form (assuming this is even possible), then you may be liable for the loss.

I’ve never heard anyone say they think twice when they send a large payment using Paymail. That they have ever actually thought to ask their wallet how large a payment is safe with Paymail. Whether it will reach the intended recipient and what the wallet will do if the recipient claims never to receive it. But these are critical details we should all know about wallets that send or receive our payments.

Unsolicited payments

The design of Paymail is structured around being able to send anyone you want money whenever you want. This is a terrible idea! If someone deposits money into your bank account and you did not expect it and do not know who it comes from, you do not touch it. You certainly do not spend it unless you are an idiot, because chances are their bank will talk to your bank and remove the money from your account. Any unsolicited payment you receive should be quarantined until you know it is safe to spend. If you are sending people money, you double check you have the right bank account number it is not a casual thing.

This is even worse with Bitcoin. People send you dust in order to tag your funds and identify where your payments go. You do not want to just include that in your available funds and spend it. Even worse still, are tainted coins. Maybe someone sends you some stolen coins and you unsuspectingly mix those into your own coins. Maybe someone sends you some coins that have gone through a mixer. There are all sorts of ways, some I am not smart enough to think of, where this could be abused.

I believe that unsolicited payments is a short term thing that seems cool, but has no longer term benefits.

Examining Paymail

Isn’t a Paymail address like email but better?

No. Email is already broken. We do not run our own email servers because it is hard and complicated. So we find ourselves in the situation where we let companies host our email and eventually we hear about people who lost access to their email address. I have vague recollections where at the first wallet workshop where Paymail was formalised, someone might have mentioned something like “Wouldn’t it be cool if Gmail supported it?” I now say it would not be cool. Imagine all your payments going to an address you no longer controlled, in much the same ways as you use that email account you can no longer access to regain access to other services which use that email account to gate-keep access.

It was also suggested at the first wallet workshop that users could host their own Paymail servers. I pointed out that no, they could not, it had been tried and people just do not have the inclination to do so. This had already been tried with OpenID where users were supposed to run their own web server to host their own identity and what happened was exactly what we saw with email, user identity centralised onto Facebook, Twitter, Google and other popular services.

Email addresses are already problematic without the link to Bitcoin, and based on the nightmares we hear about emails it is very likely tying payments to email addresses is a bad idea.

Isn’t a Paymail address just like a bank account?

No. It was suggested to me that there were modern services where you could pay to an email address, there’s even some company in the USA where they do it. You are not paying to an email address, you are paying to an account associated with that email address at some registered banking business. There are safeguards and laws that protect you as a user, either limiting liability or forcing the company to perform in certain ways to protect you. There is nothing bank-like or bank-account-like about Paymail addresses, or the businesses that host them.

We accept banks and similar licensed businesses as part of society that law knows how to deal with and that we trust. A company that runs a Paymail server is outside of that. A Paymail address is nothing like a bank account.

Without Paymail how would we do it?

In-person payment

Bob and Alice are in a restaurant. They finish their meal and then the server comes to them for a payment.

Bob and Alice get the server to split the bill and pay individually. The server holds out the restaurant’s tablet to Alice, who waves her phone over it and it contactlessly pays her portion. Bob does the same. In this case at no point is there a reason to exchange Paymail addresses, and injecting Paymail addresses into the process only serves to weirdly complicate things. The tablet likely offers a customised invoicing process to Alice’s phone, and a customised invoicing process to Bob’s phone and collects both payments to satisfy the bill.

Remote payment

Bob and Alice are in a restaurant. They finish their meal and then the server comes to them for a payment.

Alice pays for the whole meal and Bob plans to pay Alice back as he has forgotten his phone or the battery is dead. Bob has Alice in his contacts because they have a personal or professional relationship of some sort, and directly sends a payment to Alice. It is likely Alice has a peer channel open designated for contact with Bob, which allows her to receive payments whether her payment device is online or not, and pre-agreed upon templates for payment which were exchanged as part of setting up the relationship.

Personal introduction

Bob and Alice meet and decide to exchange contact details. They tap on the button on their screen, and then hold their phone out. Their phones exchange contact information. Alice knows that the entry that says “Bob” actually identifies Bob, because of the exchange. That it can be used to securely contact Bob, or perhaps even pay Bob if that ability was exchanged. Similarly Bob knows the same of Alice. And any interaction will be associated with the imported contact identity.

While it is not required, additional security can be added by a user having their identity information signed by an authority. Let’s say Tokenized offer an identity oracle, and sign Alice’s information. Then on Bob’s device given he and it recognise the trustworthiness of Tokenized’s oracle, he sees a green checkmark beside Alice’s identity. This supplements Bob’s personal knowledge that he obtained Alice’s contact information directly from her.

Indirect introduction

Carol knows Alice, and either Bob wants to get in contact with Alice or Alice wants Bob to get in contact with her. So Carol holds her phone out to Bob and shares Alice’s contact information. It might be that this is a permission Alice has to grant Carol in some way and that just because Carol has Alice’s contact information, she has no way to share it otherwise.

Remote introduction

Alice can email something to Bob — or IM or any other method of communication to existing contact points. It might be a link that Bob clicks on that opens his desktop wallet. It might be a QR code that he scans into his mobile phone. It could be anything. This can give him access to contacting Alice, or even be her contact information tailored to him.

Service sign-in

When you sign into a service, why do you have to say I am “bob@heyserviceiusethiswallet.com”? Why can’t you present an attestation from an identity oracle that gives the service the minimum information necessary (you can insert other P2P friendly mechanisms here) and just click register or login. If you think about having to provide a Paymail address to log in, all the Paymail address does is serve tie your usage of one service to your usage of another service. That does not benefit you, it benefits the two services.

All you should have to tell the service you are signing up to, is that you are a unique individual who can with some legal action against the identity oracle be held accountable. The service may ask for identifying information like your real name or KYC data, and if you think it is worth it, you can grant it access to a further attestation from the identity oracle. They might not even need access to the KYC data, but rather just the verification from an oracle they trust that the conditions are fulfilled.

Other use cases

There are many ways that external verification can be given to the correctness of the exchanged contact information — a known and trusted identity oracle can likely provide private proof for Bob from Alice using Bitcoin signatures. Or the root certificates we use in our browsers can be used to sign documents, so you can see a checkmark from “trustedoracle.com”.

Summing up

You might ask, why am I only posting about this after four years. Because discussion is unwelcome, and it is labelled as negative and even insulting to question Paymail. And to be fair it is in a sense. You do not enter someone’s house where they only eat vegetables and introduce meat, it’s just rude. But this is the way religions and political movements work, not development communities. You want to look at the confusion of people who do not get what you are promoting and developing, and address that, not attack those people for not believing in the dream.

We should at least look at what the model of using Bitcoin SV would look like without Paymail, and whether Paymail makes it better or just complicates and limits things. Whether there are fixed identifiable use cases where Paymail is the right approach and when it is not. Where is that discussion?

Please do not respond to this on Medium as I prefer not to use their comment system, discuss this with me on the Metanet.ICU Slack or the BSV Discord.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store