When should I implement multiplayer?

Should I start implementing UNET before or after "making the game" mechanic wise?

Right now i'm making a project that will be multiplayer, but I don't know if it's best to implement multiplayer now or wait and finish the main mechanics first.

 

First off: Unet will become deprecated, so you could pick some alternative. There are one or two replacements...

If you plan to do a throw-away prototype, you could first do the mechanics. Otherwise, it might actually make sense to do a few minimal multiplayer experiments and then start the project with multiplayer.

 

I agree with tobiass. Multiplayer programming is quite complex, and if you start thinking of it after your game mechanics are finished, you will be possibly forced to do some rework to make changes and adaptations to the multiplayer requirements.

 

This. For FPS/RPG/etc everything relies on spawning and unspawning things, so you end up with framework that has spawning/unspawning at it's core with all other data coming to the client for specific spawned objects.

 

It may depend on the type of game type but usually it's kind of hard to decouple the multiplayer logic from the game's mechanics. In fact typically you want to reason in a backward fashion: if your multiplayer game is also single-player you should simply emulate the multiplayer aspect with NPCs instead of actual players. So if you're making a multiplayer game you need to think about it from the get go. good luck!