Scratching My Own Itch: RollaDex

Since 2020 I've been maintaining Rolladex, an offline desktop app that I use to create and manage character sheets for playing Dungeons and Dragons. At the time of writing it boasts approximately 1.5 users: myself, and occasionally my partner. I consider it a rousing success. This post is a developer post-mortem of sorts, even though the project is technically receiving updates on a schedule of whenever I feel like it.

Why Do We Reinvent Wheels?

I started playing D&D the classic way, with the good ol' paper character sheet. It served me well for a few months, but I quickly tired of the drawbacks of physical media. It got covered in eraser markers from updated stats and hit point tallies, it was at risk of being lost or damaged by beer spills, and I regularly had to suffer the indignity of trying to parse my own godawful handwriting. On the other hand, there's a lot to like about paper character sheets. They're simple to interact with, compact to store, and information-dense, making them quick to reference.

The goal of a well-designed tool should be to keep you engaged in the activity that it supports, and not in the tool itself. When I play tabletop RPG's, I like playing adult pretend and having fun conversations about made-up adventures. I like strategizing how to best use my abilities in combat. I like reasoning about the interesting fictional worlds that the GM comes up with. What I don't like is debugging why the virtual dice roller isn't adding that extra d4 that it's supposed to in this particular scenario. I don't like haggling back and forth over the minutia of exactly what is and is not represented on the virtual tabletop grid system and why that corner is hidden by the artificial lighting system. I don't like hunting through the character building software trying to get that one feature that I know I have to be accurately represented.

Between character builders, virtual tabletops, and countless GM-centric tools, the rise in popularity of RPGs in the recent past has given birth to entire ecosystems of software that, with the goal of automating tedium, have created their own, different brand of tedium where instead of learning the rules to a system that gives you fun ways to socialize with your friends, you instead have to learn how to play a complex video game while you're also socializing with your friends.

And I absolutely get it. I'm a developer, I too feel the urge to throw code at a problem and make the coolest thing I can come up with and make my software the centerpiece of an experience. "Look at how cool this software is, look at how much stuff it does, look at how great it is to use this thing I made". For me though, RPGs are one of those lovely analog experiences where restraint is key in order to maintain a balance between giving the user power and taking the user's attention.

So, like any good developer, I thought "Surely I can do this better than everyone else", spent a weekend crapping out a basically functional prototype of what was in my head, and then proceeded to iterate on it ever since. To answer the above question in brief, we reinvent wheels because we're picky assholes who want wheels that are shaped slightly differently than all the wheels we can find on the market. Confronted with this inconvenience we can either suck it up and use what fits our desires most or chisel our own bespoke wheels, efficiency be damned. Many great, widely-beloved products and services have arisen from this model of insanity, but most people forget the many more forgotten monstrosities have been created as well.

Design Goals of My Monstrosity

In the course of a few years of following my gut, the philosophy of the app has crystallized into a few key points:

• An Offline, Desktop App Experience: No hosting fees or deployment issues, no extra strain on the network when a room full of people is already using Roll20, and characters stored the classic way, in files. Sticking your data in a web app's database is convenient in the short-term and great for the company who gets to lock you into their app's ecosystem, but worse for the long-term life of the data and your ability to control your own stuff. Since Rolladex saves to a basic plaintext .json file, I can back up my character sheets in a cloud drive, in git, on my home server, send them over email, etc.
• Long-Term Maintainability: The opposite of Resume Driven Development, I aim for boring code that still makes sense after I've taken a 6 month break and forgotten what the codebase looks like. This means minimal dependencies, low need for library updates, and a predictable project structure. This is a hobby project in support of playing games, not a side project in the "second job" sense.
• Convenient, Not Smart: RollaDex is a "paper character sheet, but better", not a fully automated D&D machine. The line between the two is more of a spectrum, and every feature I add has to navigate that spectrum and hopefully end up in the Goldilocks zone. What does this zone look like, you ask? The app will calculate your skill check bonuses, but won't tell you which ones you get to be proficient in based on your class. It allows you to hover over spells in your spell list to read their descriptions, but it's up to you to know how many spells slots you have at each level. In general, it should make moment-to-moment gameplay easier, while still putting most of the burden of knowing the rules on the user. It saves me from tedium, not from the joy of becoming proficient at something (even something as frivolous as an RPG rule system). It is emphatically not a replacement for the Player's Handbook.

This last point's approach to simplicity has two benefits. The first is selfish and lazy. The dumber the app, the less functionality I as the developer have to support. RPGs are full of quirks and edge cases that step on each other's toes, and those are sometimes easier handled by my adult human brain than meticulously mapped into code.

The second is that it enforces my game knowledge. I'd rather be forced to internalize the rules of a game than trap myself by relying on complex software to do it for me. If I'm having trouble remembering how a mechanic works, a software tool is a crutch that only solves the problem until I don't have the software anymore. If I practice using said mechanic enough, it becomes second nature and then I don't need a software crutch. The problem solves itself with time. This is probably the biggest difference between me and users of other software in this space, and also probably the biggest reason why my app would never see much popularity even if I did publicize it. "Yes I know D&D Beyond has been the default software companion for years, but come check out my app, it's got fewer features!".

Tech Stack Overview

Not being proficient with desktop UI frameworks and not being interested at the time in learning one for a prototype that might have gone nowhere, I built the UI in vanilla HTML and JS before later porting to Typescript (javascript but better) and the desktop.

• Electron: I know, I know. Electron is rightly criticized in developer circles for being a RAM-gobbling homunculus of a fake browser masquerading as an inconsistent desktop application, but the pros have outweighed the cons for me on this one. On top of being able to get up and running with web tech, fast (pushing out shady javascript and fixing it later was pretty much how I started my career as a developer), the cross-platform nature has come in clutch. When I started this project I was running desktop Linux, but now use a Macbook. My partner uses Windows. If I had needed significant effort to port the codebase to all of those places, I likely would have abandoned the project long ago. That being said, Electron is far from the only crossplat name in the game these days, and I would look into more performance-friendly alternatives if I were to ever start from scratch with a V2 (unlikely, but not out of the question if I ever get tragically bored in my spare time).
• Knockout: Knockout is practically Bronze Age tech in Software Years, but I wanted a plug-and-play library that would help me with databinding without the weight of a more opinionated framework. Knockout fit the bill, and has been relatively easy to maintain while minimally dictating the structure of the codebase. That being said, the longer the project goes on the further it moves along the sliding scale from Glorified Web Form to full-blown Web Application, and a more modern framework that prioritizes reusable components and a SPA philosophy would be the first thing I'd explore in a hypothetical re-do (assuming I was still using desktop web tech at all).
• Shoutout to RPGUI: When the project finally left the programmer-designer hallmark stage of default Bootstrap styles, I found RPGUI, a neat little css library that gives the app a retro video game feel that fits old school fantasy very well. It even inspired me to make a little landing page with animated torch gifs and a menu for choosing from recently opened characters. Layout-wise, the app is still just a copy of the paper character sheet layout mapped more or less directly onto HTML. I may one day decide to redo the layout from first principles without relying on what Wizard of the Coast designed, but for now it's functional enough that I haven't felt the need.

Future of the App

RollaDex is in a more-or-less permanent state of being unfinished. That being said, I've been using it as my full-time character sheet solution ever since that first prototype flopped out the door. Every few months or so I'll circle back to add a new feature or make a small fix, but the pace has slowed as of late. My group has recently started experimenting more with Pathfinder 2e, and the next large job would probably be to flesh out the currently half-assed support for that game's character sheets. If this post interests you then feel free to check out the project on Github, but don't expect a ton of attention if you submit an issue. I'm not interested in turning this occasional side pursuit into a significant time-sink where I have to support user requests or manage community contributions. Despite being open-source because Well Why the Hell Shouldn't I?, RollaDex is a home-cooked meal. You're welcome to have a peek or a share or even just some inspiration if you'd like, but at the end of the day it's mine. Small on purpose, and exclusively the way I like it.