Building a graphical multi-user spreadsheet editor in Zig, introduction

Note: this article is part of the "Building a graphical multi-user spreadsheet editor in Zig" series. Read all the articles here:

series homepage

It's been a while since my last project (edit, a text editor in C). I've had many other project ideas since then, but I haven't taken the time to bring them to life. Now I have some free time over the next two months, and I'm willing to use some of it for a nice software project: grid, a spreadsheet editor!

Why I'm writing about it

This blog entry is (hopefully!) the first of a series documenting my progress. There are two reasons to do that.

First, I expect to explore many new topics during the project, wether it is Zig features, computer science concepts, data structures, or algorithms. Writing about my findings will allow me to clarify my understanding, and you might find them interesting as well.

Second, it might drive my motivation to carry on the project, as I will want to wrap this series up nicely.

Why a spreadsheet editor

Unlike my text editor, I very seldom use spreadsheet editors. I think it's a key point in the project: I'm not designing and building grid for my personal use, frustrated with existing tools, but for the learning experience. Thus, what do I expect to learn?

• Multi-user software. Latencies between users induce state replicas inconsistencies, which calls for consistency restoration and data integrity preservation tools.

• Networking. grid will use a client-server model, communicating through sockets with yet-to-be-designed protocols. If it ends up on a publicly available web "application", it will require security stuff.

• Multithreading, for both the server (managing several clients) and the client (modifying the state considering user actions, rendering to the screen, handling user input, handling the connection with the server).

• Graphical interface. raylib caught my eye recently. Even though I keep the idea of a text-based interface in mind, leveraging the information display density of the graphical interface seems to be the good fit for something as complex as spreasheets.

raylib

• Demanding interpretation engine and data management. Spreadsheets cells contain formulas to parse and compute (and cells might depend on others cells). Spreadsheets can hold large amounts of data. Managing all that in a high-performance way calls for optimized data structures and some graph theory.

Quite an ambitious roadmap!

Why Zig

I've never used Zig. However, I figured it might be a good fit for this project.

Zig

First, Zig is advertised as a simple language, something I like and need for performance. Its great C-interoperability should allow me to use the unmodified raylib source code (in C), and to get back to C for some parts.

In addition, its cross-compilation abilities enable targeting several client types: native (any platform) or website (with WebAssembly). As I develop on WSL Linux but can only run graphical applications on Windows, cross compilation is especially precious.

Finally, I'm curious of Zig modern (compared to C) and innovative features: payload capture, tests, struct methods, optionals, defer, error management...

Goals

My goals for grid are:

• to implement the points of the "Why a spreadsheet editor" section as well as possible

• to design a capable data model that works well with text-based version control systems

• to offer a responsive (no matter how good the connection to the server is) and explicit (easy to explore features and to understand what's going on) user experience

And aside from the final product:

• to write idiomatic Zig code

• to post frequent updates on the evolution of the project

• not to prematurely optimize