NFTs, explained by a software developer

In this article, you will understand what an NFT is and how we developers can create one ( or maybe 10,000).

An NFT is a transaction recorded on a blockchain that attests the ownership of a unique object, linked via metadata to resources, usually images or video.

Okay, this definition is so general that it doesn't explain anything when you read it like this.

Let's take a step back and start with the basics.

shopping cart filled with items

Photo by Ananthu Ganesh on Unsplash

What does fungible mean?

Imagine you are at the supermarket, and you get a coupon or stamp for every $30 you spend. You are a loyal customer and have 20 half-dollar coupons.

Now imagine that as you rush out the doors with your coupons in hand, a middle-aged lady bumps into you, and your 20 coupons fall out, mingling with the lady's ones.

No big deal, you know you had 20, and the lady knew she had 10. You pick them up off the ground and share them. Each one is fine.

If we think of coupons as tokens, they are fungible: you can exchange them for each other. Each has the same value and characteristics as the other.

If you dropped your house keys instead of coupons, that was different.

Although visually similar, each key differs from the other, and you could not interchange them with the lady. If we think of keys as tokens, these are non-fungible tokens.

> Student: So, are house keys NFTs? 
> Guru: No, because they do not "live" on the blockchain. 
> Student: Right, and how does the blockchain work?

NFTs on the blockchain

Now that we know what non-fungible means, let's figure out how to connect these tokens to the blockchain.

The good news is that it is optional to know in detail how a blockchain works to develop an NFT. We need to know some basic information. Let's see it.

The first blockchain - Bitcoin - was born in 2008 by Satoshi Nakamoto: however, the subsequent implementations are all based on some standard features.

The blockchain (at its base) is a distributed data ledger, i.e., we can imagine it as a database or data storage in which transactions, i.e., changes in the state of the blockchain, are written.

Its fundamental characteristic is that it is not a single, centralized data source. Still, it is distributed in a peer-to-peer network, where each computer (called a node) has a copy of the entire ledger.

The addition of data is done through consensus mechanisms because most of the network nodes validate the transaction.

The essential characteristics of the blockchain are:

it is distributed (precisely because it is decentralized).
it is transparent (because the ledger is public and open).
it is trustless (that is, it has no central authority or intermediaries that users must rely on, but anyone can operate directly on the network)
it is incorruptible (the data written on the blockchain is permanent and immutable)
it is secure (because it relies on mechanisms that ensure that data are validated, using asymmetric key cryptography and various hashing systems)

A key feature that gives the technology its name is the way the data are stored: they are grouped in a block structure concatenated together in chronological order.

Writing data to the blockchain is done by generating a new block containing a certain number of records (named transactions).

The current block's hash is calculated, then inserted as a reference within the next block, thus concatenating the two blocks. As new blocks are added, the chain grows.

Getting closer to NFTs

> student: okay, but you still need to tell me how to develop an NFT.
> guru: Easy, you can't run before you know how to walk.

In the Bitcoin blockchain, transactions essentially contain currency transfers from one account to another. Still, in 2015, thanks to the work of Vitalik Buterin, Ethereum was born.

This new blockchain allows code to run on it. This code, these programs that run on the blockchain, are called smart contracts.

Each node on the Ethereum network implements a system called the Ethereum Virtual Machine (EVM), a virtual machine capable of executing smart contracts code, using the local copy of the blockchain to read and write information, and possibly interacting with other smart contracts.

Finally, NFTs

NFTs are a particular type of smart contract that implement a specific standard. This standard is ERC-721, which defines an interface to be implemented by the smart contract.

This interface has functions for determining the ownership of an NFT and functions for transferring it.

Then there are functions to obtain the name and symbol of the NFT.

An essential function is tokenURI, which defines a URI for each of the tokens managed by the contract.

This URI must point to a JSON file with a structure like the one defined in the figure. The standard defines the first three fields. The attributes are an extension added by the OpenSea marketplace and then became very common.

The image attribute points to the actual image.

Creating NFTs

Typically, users interact with a smart contract through a dApp (decentralized application), which interacts with the smart contracts deployed on the blockchain.

Usually, the dApp takes care of getting the user connected through its wallet. At this point, it allows the user to perform minting, that is: generating a new NFT from the smart contract.

The contract will write a transaction on the blockchain, generating a new token with its associated token id.

This token id is used to retrieve the metadata and image (usually previously uploaded to a distributed filesystem, e.g., IPFS) to show the image of the newly acquired NFT to the user.

At this point, we can write the initial definition again, which I hope has become more evident:

An NFT is a transaction recorded on a blockchain that attests the ownership of a unique object, linked via metadata to resources, usually images or video.

> student: now I understand, but in practice how should I develop an NFT?

Yes, but in practice?

To develop the smart contract code, you use a programming language called Solidity.

The ERC-721 interface has already been implemented for us by Openzeppelin, a company that has written many open-source smart contracts.

Some frameworks and tools help us develop the smart contract of an NFT, test it, deploy it on the blockchain, interact with the deployed contract, and finally create a dApp for users to interact with it.

Among the most popular are truffle and hardhat.

To develop an NFT, you should follow these steps:

Generate the assets (images, video, audio, etc.) and upload them online. Preferably on a distributed file system such as IPFS.
Copy the URI of the assets.
Develop the smart contract that implements ERC-721. You can use Openzeppelin's ERC-721 basic smart contract to make your life easier.
Write tests on the newly developed smart contract. Once on the blockchain, the smart contract is not editable: our bugs will remain on the blockchain forever.
Deploy the smart contract on a test network. Yes, testing environments exist in the blockchain as well, fortunately.
Run tests of mint and other transactions on a test network.
Deploy the smart contract on a production network. This step requires hooking into a node on the blockchain.
Optional but recommended: develop a dApp to allow users to create NFTs (usually paying a fee).

> Student: but how do I connect to a blockchain node?

Bonus: deploy the contract on the blockchain

Connecting to a blockchain node is a challenging task.

But first, let's see how a node works. Each blockchain node consists of essential parts:

the blockchain data
the EVM (Ethereum Virtual Machine) to execute smart contracts
JSON-RPC APIs that the node exposes to allow users and other nodes to communicate with it.

When a user wants to interact with the blockchain, he must connect to a node, and he does so by calling the APIs that each node exposes.

To do so, he needs to know the Internet address of this API.

Moreover, these APIs are not the classic REST APIs we are used to using every day but are "slightly" more complex. For two reasons:

1. the content of the messages, which follow a standard, are encoded in hexadecimal

2. You must sign most of the messages with your private key.

You should use a ready-made library to implement calls to a blockchain node. There are some for almost every language.

But that's not the end of the story. It gets even more complicated, the nodes may need to be in perfect sync with the rest of the blockchain, or it may be turned off or become unreachable just as it handles one of our transactions.

How do we solve this problem?

> Student: why can't I launch a blockchain node by myself?

            In the end, it's just a program running on a server.

> Guru: Running a blockchain node on your pc or server is not trivial.

        Server space in the hard disk (about a terabyte currently), a lot of RAM, CPU, 
        and download internet bandwidth. 
        You also need to synchronize all existing blockchain data before you

can start receiving calls to "run smart contracts."

Fortunately, there are so-called providers, i.e., services that expose the same API as the blockchain nodes but are more reliable and dedicated to these operations. The most famous providers are certainly Infura and Alchemy.

So, when you want to deploy the NFT's smart contract on the blockchain, you must connect to a provider and perform a "special" contract creation transaction.

When the contract gets deployed on the blockchain, it assigns an address to it.

The blockchain contract address is unique and must be used to execute the various transactions that the smart contract exposes: for example, the mint or transfer of an NFT.

Closing up

We have given an overview of what it takes to create an NFT. We have barely scratched the surface of this world, and in this article, there was no way to get into the depths of development details, but stay tuned. I will post more implementation-specific material as soon as possible.

In the meantime, if you want to learn how to develop an NFT or smart contract, feel free to contact me.

Thanks, and until next time