Ara: Making smart contract transactions smarter than ever

• Real-time transaction information now accessible
• Transactions can be securely held between parties until specified conditions are fulfilled
• Includes new tool to build and train neural networks in Etch

Introduction

Fetch.ai has taken another step towards the release of its mainnet in December. Our latest technical release, Ara, includes major new updates to transaction functionality and visibility.

Context is key

Developers using Etch, our unique smart contract language, now have access to a ledger context object. This has access to transaction information such as the block number the transaction is in, and much more. Using this context object, developers can also view the details of the transaction. This makes it possible to check at the smart contract level who has signed the contract, how much has been transferred, how old the contract is, how long it will be valid, as well as many other details. It enables a vast number of real-world opportunities, including validation by groups of signatories. For example, if a payment to Person A needs the approval of both Person B and Person C, the smart contract can see who has signed the contract and determine whether the payment should be made.

The balance of power

Another example is that money/assets/services linked to a smart contract can now be kept in escrow. To those unfamiliar with the term, this means money/assets/services which are set to be involved in a transaction can be kept in the custody of a secure third party. Today’s technical release means that goods held in escrow on the Fetch.ai network are withheld by the smart contract and only exchange hands when a condition, specified at the time of the creation of the smart contract, is fulfilled. To enable this, and many other features, a smart contract can now transfer FET to receiving entities and check the FET balance.

Another string to the developer’s bow

This release also offers developers a new and exciting Machine Learning and AI tool in the Etch language. It is the simplest way to build, train, and evaluate neural networks in Etch. The tool takes care of the underlying implementation details for Graph, DataLoader, and Optimiser, allowing you to build complex training applications with minimum code. Check out our cool example.

AEA documentation now live!

The documentation site for the Autonomous Economic Agents Framework is now live. Take a look.

Get involved

Thanks to your input, we have fixed a number of bugs in our Aquila ledger release. We would really appreciate your help to review Ara. Don’t forget, as part of our Technical Bounty program, we are awarding up to $10,000 for critical issues reported in our GitHub ledger repository.

Coming up

The launch of the Fetch.ai mainnet at the end of the year is fast approaching and we’re hugely excited by the journey ahead.

For more information, check out our supporting documentation or join our developer Slack channel.