Where you are. You have the whole legacy machine. Five modules built it: money at rest on a two-tier ledger, domestic rails, the crossing of borders, the markets that price things, and the leverage and clearing that keep those markets standing. Every one of those ran on the miniledger you wrote in module 1. This module keeps the miniledger and changes the shape of the ledger underneath it.
Alice pays Bob a hundred pounds. She banks at Alder, he banks at Birch.
You know exactly what happens now, because you built it. Alder debits Alice and credits its own settlement obligation. The central bank moves reserves from Alder to Birch. Birch credits Bob. Three institutions, three books, and a sequence that has to run in order, because Birch will not credit Bob against a promise it has no reason to trust.
Now suppose Alice and Bob both keep their accounts in the same book, and the book is not owned by anybody in particular.
The idea in one paragraph
A blockchain, for the purposes of this course, is one shared append-only ledger: many machines hold identical copies of the same list, and they run an agreement procedure so that every copy takes the same next entry. There are no tiers, because there is only one book; there is no correspondent, because there is nobody to relay through; and there is no single bookkeeper, because no one machine’s copy is the authoritative one. The price of that arrangement is that changing the list costs something real, since every machine must do the work of agreeing. What you get for the price is that a transfer between two accounts on the list is one posting rather than a sequence of promises between institutions that settle later.
Who is allowed to hold the pen
There is a second question, separate from the ledger’s shape, and mixing the two is the most common confusion in this area.
The first question is what the ledger looks like: one list or many. The second is who may write to it. A public ledger lets anyone run a copy and submit entries. A permissioned one restricts that to a known set of institutions, which is what most bank-led projects in this space actually run.
The two questions are independent, and the answers get combined in all four ways in practice. A course that treats “blockchain” as one thing cannot explain why a bank consortium and a public network make opposite choices about the same trade-off.
Wider than the screen; scroll it sideways.
What the exercise counts
The exercise runs the same payment both ways and counts changed ledger lines.
On the two-tier model, six lines change: Alice’s deposit, Alder’s reserve balance, Alder’s obligation, Birch’s reserve balance, Birch’s obligation, Bob’s deposit. On the shared ledger, two lines change: Alice’s balance and Bob’s.
That number is the whole claim of this module reduced to something you can count. It is not an argument that six is bad; module 2 spent a great deal of effort explaining why those six exist and what each one is protecting against. It is a statement that the two arrangements are genuinely different, and that the difference is measurable rather than rhetorical.
What the rest of the module does with this
Everything from here is built on that single shared book. The next seven lessons put a dollar on it and stress the result until it breaks. Then the module widens out: deposits and central bank money in token form, and assets that are not money at all.
The one thing to carry forward is that the shared ledger is a container. It does not tell you what a token on it is worth, who owes you, or what happens when that party fails. Those questions have the same answers they had in module 1, and this module keeps asking them.
Check yourself
1. What is the defining property of a shared ledger, in this course’s terms?
That there is one book rather than tiers of them. Many machines hold identical copies and agree on each entry, so a transfer between two accounts on the ledger is a single posting instead of a relay of promises between separately owned institutions.
2. Alice pays Bob across two banks: six ledger lines change. On one shared ledger the same payment changes two. What does that difference actually measure?
The number of separately owned books the payment has to cross. It measures structure, not speed or cost. A shared ledger can be slower and dearer than a domestic rail; the count says only that the coordination between institutions has been removed, because there are no longer several institutions to coordinate.
3. Why are “one shared ledger” and “anyone may write to it” separate questions?
Because the ledger’s shape and its access rules are independent choices, and all four combinations exist. Bank consortia typically want the single-book shape with a vetted set of writers; public networks want open participation. Treating “blockchain” as a single design cannot explain why the two make opposite choices.
4. What does a shared ledger not tell you about a token sitting on it?
What it is worth, who owes it to you, and what happens to you if that party fails. Those are questions about the claim, not the container, and they have exactly the answers module 1 gave them. The rest of this module is mostly about asking them carefully.
Do this
Post one payment two ways and count.
python3 code/one_ledger.pyThe starter gives you the two-tier run using the World you built in module 1, and leaves you shared_ledger_hops: stand up a single Ledger, open Alice and Bob on it, post the same hundred, and count the lines that changed. The assertions check that the two-tier path changes six and the shared path changes two.
The completed version is in solutions/one_ledger.py.
What you can now do. You can describe a shared ledger without reaching for the word blockchain, separate the question of the ledger’s shape from the question of who may write to it, and measure the structural difference between the two arrangements in changed lines. The next lesson puts a dollar on the shared ledger and asks who owes it to you.