Polkadot vs Rococo runtimes

Question

I noticed various projects having multiple runtimes for Kusama, Polkadot and maybe a local-testnet. This makes sense as they can have different set of features.

In my case I only want to test on Rococo and launch on Polkadot. Is it recommended that I have separate runtimes for each? If so what are reasons in having multiple runtimes? Is there a recommended setup for this?

Do Rococo and Polkadot have the same versions?

In the documentation it states:

Parachains are very tightly coupled with the relay chain codebase that they connect to because they share so many common dependencies.

Answer 1

First of all, Rococo has never meant to be a canary network for any other relaychain. And its value and appeal resides in having a testing hub with the maximum possible number of ecosystem networks present on it, enabling testing intercommunication and integration between different chains among other things.

Though, Rococo runtime, spec, config and such are not exactly the same than the ones running in Kusama or Polkadot but it tries to mirror the same set of features that we can find in Kusama. Its runtime gets released at every Polkadot release. Check this release for instance, you will find Rococo's runtime being released too. So,

Do Rococo and Polkadot have the same versions?

Both runtimes are released together, with the same version.

Is it recommended that I have separate runtimes for each?

For this question I would recommend what I have been observing among the different teams using Rococo network at the moment. Yes separate runtimes seems the way to go, though these runtimes are effectively the same one, adding specific values for Rococo in your spec, for example

Answer 2

You should have separate runtimes for the Testnet and Mainnet. That way, you have a testing ground that you can experiment on without any risks of dire economic consequences.