Stop Killing Games

The obvious case where this makes sense are single-player games that require internet access before they even launch, like when you need to link a Microsoft account to play Forza.

But it's less obvious to me how the legislation should work for a multiplayer-only game that goes out of business. I suppose it should require a refund at some point. But at what point?

Steam only lets you refund a game that you played for less than two hours.

And if you think that's not long enough, there's surely some time period where you can agree that you've got your money's worth. Kind of like how you lose the ability to say "I didn't like it" after you ate your whole dinner at a restaurant.

Yet in the comments here someone gives an example of three years of online support which is insane. Why is multiplayer special? Should Steam also let you refund any game until three years elapse?

> But it's less obvious to me how the legislation should work for a multiplayer-only game that goes out of business

The idea is to force the companies to provide an end-of-life plan that leaves the game in a "reasonably playable state". Depending on the game the approach can differ. Releasing the binaries along with instructions for hosting dedicated servers is one approach (that some games already take). I was playing on a (pirated, but it was a long way ago and I was a child) Lineage 2 server way back in 2007 on my local ISP network, so even something like MMORPGs can be covered if it's included in the discussion at the design stage of the game.

But even if the backend is very complex and vendor-locked, releasing something like a set of Cloudformation templates and saying "you can only host this on AWS and it's going to be restrictively expensive but there you go" is also an option that would satisfy the requirements. It's still better then having nothing at all to dig at (although some fans still do reverse-engineer and spin up community servers even without having access to any of this).

The problem is that you can easily create negative effects that hurt all developers while only imagining the impact on profitable AAA games which seems to be how most of these discussions go since people always bring up blockbuster games like Lineage 2 and World of Warcraft.

Ensuring that the multiplayer server component of your game is a standalone end-user distributable is a huge task to impose on every game that wants to have a multiplayer component. Especially once you consider the vast majority of games that never even get traction much less turn a profit.

So, the second someone buys your prerelease indie slither.io game, what exactly does this checklist look like? It needs to also day 1 launch with a self-hostable standalone server distro instead of the crappy spaghetti mess you live coded on an EC2 machine?

The requirement is not to be standalone end-user distributable, the requirement is that it be somehow possible for an end user to set it up, which is a lot easier, unless the developers don't even know how to set up the backend. But that's a low bar still.

I've built a couple multiplayer browser games and getting the server component to a state where it's end-user distributable probably involves another 50% of the work that I put into building it in the first place since I'm just rsyncing code onto a VPS.

It's like thinking that just because the code exists, then it's in a state that could be pushed publicly to github, and that's not the case for almost any codebase.

To think that I would need to do all that the second I charge $1 seems unreasonable. And I think you underestimate how true this is for most games you see on Steam.