Using Obsidian on a Blockchain¶

This tutorial will not try to teach you about blockchains; there are plenty of other tutorials for that. For now, you only need to know that blockchains are server-based systems that maintain state safely. The blockchain system processes transaction invocations, which come from clients. Your focus so far has been on implementing applications that can run on servers.

To run transactions, then, we need both a client and the blockchain:

First, instantiate a contract on a blockchain.

Then, clients can invoke transactions that were defined by the contract.

To instantiate a Obsidian contract directly on the blockchain, it must be a main contract. Because every client must have a reference to the contract, all the contract’s transactions must have this be a Shared reference. But what if the contract is an asset? As a special exception to the usual rules, the blockchain itself is considered the owner of the instance for the purpose of avoiding asset loss. For example:

main asset contract Bank {
   Money@Owned vault;

   transaction deposit(Bank@Shared this, Money@Owned >> Unowned money) {
      // put money in the vault...
      // (code not shown)
   }
}

If the Bank contract is instantiated on the blockchain, the blockchain itself ensures that the Bank is never lost, and clients can execute transactions via their Shared references to the Bank.

If you need to understand how clients work, you can refer to Obsidian Clients, but that is not necessary on your first read of this tutorial.

Concurrency¶

On the blockchain, all transactions execute sequentially. As a result, you do not have to worry about concurrency in Obsidian; if your transaction is executing, no other transactions are executing concurrently.