So, now let's talk about how
Bitcoin blocks get put together.
So we've been talking
about the blockchain and
the fact that transactions
are actually grouped by blocks.
But everything in this lecture so far,
we've been talking about just individual
transactions getting published.
So why do we group transactions
together into blocks?
Well, a couple reasons.
One is that that creates a nice,
single unit of work for
miners that's bigger that
the individual transaction size.
So if the miners had to do work and
do hashing and add metadata for
every transaction in the system,
that would provide too much overhead.
It also makes the hash
chain of blocks shorter
because we only need one block for
a large number of transactions.
And that's going to make it easier for us
to verify the blockchain data structure.
So what does the blockchain
data structure look like?
It's a pretty clever combination of two
different hash-based data structures.
So on the top here,
we have a hash chain of blocks.
Each one has a block header and
then a pointer to some transaction data,
as well as pointer to the previous
block in the sequence.
And remember these are hash pointers.
And then we have a tree of all of
the transactions that are included in
each block.
So this is a hash tree or
what's called a Merkle tree.
Which commits to all of the transactions
in the block in quite an efficient way.
So it's easy to provide just the path
through the tree which will be logarithmic
in size to prove that that transaction
is included in a specific block.
So that's the high level
idea behind the block.
What does it look like in
practice at the low level?
So we'll do another deep dive
into the actual data here.
So this is what a Bitcoin
block looks like.
There's what's called the block header
which has all of the metadata for
that block.
And then there's that Merkel
tree of transactions.
So basically a long list of transactions.
All of the hashes are arranged in
this tree structure, which gives you
the ability to efficiently prove which
transactions are included in a block.
The most important part, of course, is
the header, which mostly has information
related to the mining puzzle, which was
talked about in the lecture on consensus.
And we'll revisit in
the lecture on mining.
But recall that the most important thing
here is that the hash of the block header
has to start with a large number of
zeros for the block to be valid.
And then there's some other
data to make that happen.
There's a nonce that miners can change.
There's a time stamp.
There's an indication of how
difficult this block was to find.
That's all stored in the header.
And the important thing is that
the header is the only thing that's
hashed during mining.
So to verify that chain of blocks,
all you need to do is look at the headers.
And the only transaction data
that's included in the header
is that one route of the transaction tree.
So that's this mrkl_root parameter.
The other thing that's interesting about
blocks is that they have one special
transaction in the Merkle tree which
is unlike all the other transactions
we looked at before.
And this is the coin-based transaction.
So this is where the creation
of new coins in Bitcoin happens.
It mostly looks like a normal
transaction with a few exceptions.
So the value of this transaction is
going to be equal to currently a little
over 25 Bitcoins.
As we discussed, this is a flat mining
reward which is set by the system and
which is halving every four years.
In practice it will be a little
bit more than 25 Bitcoin,
because it also gets to include
the transaction fees collected from
every transaction included in the block.
So the pointer to the output transaction
that this coinbase transaction is
receiving is a null pointer.
It's a hash of all zeroes.
And this is an indication that since
this is the creation of new coins,
there is no antecedent.
There's no previous transaction that's
being consumed to create these coins.
And there's also this special
coinbase parameter, and
the coinbase parameter
is completely arbitrary.
The miners can put whatever
they want in there.
So, famously,
in the very block ever mined in Bitcoin,
the coinbase parameter had
a quote from the newspaper.
It had a quote from the Times of
London describing a story involving
the chancellor bailing out banks,
which was both a political commentary on
the motivation for starting Bitcoin, and
served as a commitment that Bitcoin,
the first block was obviously mined
after this newspaper came out.
But since then,
miners are free to put whatever they
want in the coinbase parameter.
It's been used as a place to put some
arbitrary data for different reasons.
To signal support by miners for
new features, but
there's no actual limits on
what miners can put in there.
So with the block format and
the transaction format that we described
earlier, the best way to learn
it is to just see for yourself.
So there's a lot of websites
that make this data accessible.
This is a screenshot from blockchain.info
which I found very helpful
myself in understanding and exploring
what's going on at a low level in Bitcoin.
There are a lot of other websites that do
a great job of making this information
accessible.
So you can look at
the graph of transactions,
see which transactions redeem
which other transactions.
Look for
transactions with complicated scripts,
even look at the block structure and
see how blocks refer to other blocks.
It's all available online, because again,
the Bitcoin is a public data structure, so
a lot of different people
have put very pretty wrappers
around this to explore it graphically.