The Curse of Good Enough

There are lots of tools out there, physical and digital, that are really bad that are nonetheless still in use. Often it’s a case that the people who use the tools don’t have any choice, either because the tools are too complicated for them to reasonable create their own or some more complicated reason they can’t (regulatory reasons for example). But I feel quite strongly that there is a second category of tools that are still bad (and in some cases worse), but they’re just useful enough that they suck out any effort that could be spent on improving them.

It’s no secret that innovation takes time, but there’s a whole category of tools that enable their users to do most of what they want, most of the time and are just awful enough to use that they don’t leave their most invested users any time to improve. A lot of people who work on tooling will have seen a scenario like this, where putting out the fires the use of a tool makes takes up any available effort to actually stop the tool from starting fires.

Specifically here, I want to talk about a tool called Tabletop Simulator. It’s a $20 piece of software that provides a vaguely boardgame themed physics sandbox in a multiplayer environment. It’s all the fun of Gary’s Mod, but with primatives for a deck of cards instead of posing TF2 models with anime characters. It is, to be fair, an amazingly powerful piece of software that has enabled a huge number of people to test and play a variety of tabletop oriented games. It’s also quite janky.

To say it’s got rough edges is an understatement. The authorial experience with it is slow and tedious. It’s got an API but I’ve never seen anyone speak fondly of it. The physics often gets in the way and visual glitches aren’t uncommon. It’s also got more than a few baked in assumptions about how games are supposed to work and these aren’t always obvious.

That being said, if what you want to do fits how Tabletop Simulator works, it’s a great tool. The problem is that once you’ve started it’s all too easy to fall foul of the sunk cost fallacy. Is it better to sink ever increasing amounts of effort into working around it’s issues (it can be extended both through Unity assetbundles and Lua scripting) than to start again with a more suitable tool? Many of the options one might take are also bound by the issue whereby it’s damn near impossible to share such improvements generically. It being closed source software isn’t strictly bad, but I don’t feel like there’s a massive level of improvement coming from the developers and the model means there’s not a huge amount coming from the community either.

There are alternatives. Tabletopia and Board Game Arena are both browser based competitors, although Board Game Arena looks to be as accessible as writing your own game from scratch to your average board game designer and Tabletopia charges $20 a month for a comparable feature set to Tabletop Simulator and still only allows a limited number of games. Being centralised websites, you’re also stuffed if they ever go offline. I could (in theory) backup a copy of Tabletop Simulator and keep a single golden version forever (although it does rely quite heavily on its integration with Steam, but you can work around that).

I lost several hours this morning to a stupid error message so I’m feeling quite salty. But it did get me thinking, there has to be an alternative to this.