Introduction

In this chapter we set the stage, clarifying who this book is for, how to read it, and to expect.

Who this book is for

We assume that we already have some basic Rust coding skills: for example, what are traits and how to implement them for our types; what are generic types, and how to handle results representing either success or failures. We do not assume any prior knowledge of Serde internals, neither we assume knowledge of the UBJ - Universal Binary JSON format. We introduce concepts from first principles, and we grow the complexity step by step.

How to read this book

We recommend reading the chapters in order. The first three chapters set up the problem and the architecture. The two middle parts build the format layer and then the Serde layer, mirroring the order in which the real crate is written. The final part covers the practical concerns that make the implementation usable in the real world.

What we are going to build

By the end of the book we will understand, layer by layer, how a Serde format can be implemented in Rust language. We will be implementing a Rust library crate, named serde-ubj, targeting one particular data format, named UBJ - Universal Binary JSON (see https://ubjson.org/).

Our Rust crate is going to be layered as follows:

• the wire format
  the binary markers and payloads that define UBJ on disk or on the network, as defined by the official format specification (see );

• the format layer
  small reader and writer Rust types that know the “on-the-wire” bytes but nothing about Serde;

• the Serde layer
  the Serializer/Deserializer adapters that translate the Serde data model into/from those bytes;

• the edge cases
  enums, options, no_std support, buffering, and testing.