A Design View of blockchains
To create a web of money, you need two way links so that the numbers add up when money moves from one place to another. This is the fundamental UX concept underlying cryptocurrency protocols.
Blockchain Use Case & Architecture and the History of the Web
In a previous post, I argued that the primary use case of blockchains was security, and decentralisation was the means to achieve it, rather than the other way around, which is the more common way of looking at them. The argument was based on the history of the web and the previous wave of decentralisation in 2001.
This post is about the fundamental design architecture of blockchains. I’ll argue that they are about two-way links which also relates to the history of the web.
One -Way Permissionless Linking Created the Information Internet
The Internet as a communications medium predates it as an information one, and email was perhaps the most common example of an Internet communications protocol (SMTP) that became popular before the the web (HTTP). The fact that HTTP has evolved to subsume email client interfaces within it, often makes people forget that email is based on a different underlying protocol.
The web was created as an information medium not a messaging one and this meant being able to publish and read (browse) articles. There were rival protocols such as Archie, Wais and most notably Gopher, but these didn’t get the fundamental architectural design right for them to take off. What made the web work was the idea of one-way links, links to any ‘page’ from within a portion of another one — an idea called hypertext.
It was very important to have one way links for the Internet to grow, as it meant anyone could publish something new and link it to an existing web page to become part of the same network, without asking the permission of the other publisher. This concept is something that traditional publishers still often fight against, today. If the links were two way, that permission would have been required.
Two-Way Links for the Web
In the early days of blogging, Six Apart, creators of Moveable Type baked in permission for return links by adding the ability to automatically add the link back from an inbound one by looking at the server logs. This ingenious method, called a Trackback was the first example of widespread two-way links on the web.
By pre-baking the permission in, the spread of two-way links was hampered by the fact the were prone to spam. They had the opposite effect of the two way permissioned linking envisaged by newspaper groups who lobbied to to be removed from Google News so they could have absolute control. The spam often occurred in comments on websites with large numbers of pre-existing inbound links, and therefore high Google PageRank, which meant that the auto-triggered return links were valuable.
The economics and incentives of the web mean that one-way linking remains the dominant model.
Two-Way Permissionless Linking Has Created the Internet of Money
For an Internet of money, two-way links are essential because an internet of money has to show transactions and balances and each transaction changes balances in two places. This is the foundation of modern accounting and double entry bookkeeping, however ‘books’ deal with transactions on a single balance sheet, whereas a web of money would have to deal with multiple balance sheets i.e. be multiple linked ledgers.
Two-way Permissionless Links for Transactions Logically Imply a Universal Shared Ledger
In order to be able to enter transactions on one ledger and have a linked one belonging to someone else update without permission, you need not just linked ledgers, but a universal one— ie everyone shares the same transaction history. This logic shows, from a design perspective, why an Internet of money has to be a shared ledger and how this can only be achieved via permissionless two-way links (i.e. transaction updates from a ledger representing both your own balance sheet and someone else’s).
A Universal Shared Ledger Solves the Double Spend Problem
From this logic, which argues that a universal ledger is the only way a web of money could have ever been, the other advantages of this model are suddenly possible, such as the solution to the double spend problem for digital currencies.
An inherent problem with digital goods is that they are the opposite of scarce, digital goods are just numbers and can be copied for almost nothing. They are forgeable. This means that the only way to make numbers unique is to have a database of ownership where one number can only belong to another number representing an address.
One person can allocate the addresses or, in the case of blockchains, the network can, by consensus. By making it impossible to allocate changes to addresses except by consensus, the network doesn’t have one point of failure and so is resilient to attack — something which is particularly important if your database is money.
In the same way that the web allowed anyone to put a new publication on the network, without having to go through a publisher, a web of money allows anyone to add a balance sheet without going through a bank. Clearly that is important.
Bi-directional Transaction Links ‘Hypercash’, are the Foundation Stone of the Financial Services Internet.
From the concept of of bi-directional links; a mechanism for anyone to create them and a universal network, owned by nobody, you have a structural model which represents both public blockchains and an Internet of Money. It’s possible that these protocols really will become that, and fundamentally because they use the concept of bi-directional links rather than the uni-directional ones that the web is based on.