Is rococo-parachain necessary?

[NachoPal]

Coming from: #1223

I would like to understand what is the role of rococo-parachain. I think parachain-template should be sufficient and the way to go if you want to tests/tinker/build a parachain.

It seems to me that having both runtimes coexisting is redundant. rococo-parachain configuration is closer to what we (the Common Goods parachains team) want to achieve (a parachain template ready to interact with a Relay chain and its Common Good parachains), however I think that logic should go to the template. Furthermore, right now it is unclear to me what configuration parachain-template is following.

I would like to propose removing rococo-parachain unless there is a reason for keeping both runtimes. It will be also one runtime less to maintain.

[nuke-web3]

#1223 (comment)

This template is used to start most parachain projects from, and thus is intended to be minimal AKAIU, thus adding more functionality here, especially divergent from the substrate template, isn't ideal.

If common good chains need a more purpose built template, I would think it should live independent from the soverign template (we have now) if significantly different. Especially in the XCM configuration, because this template should never be set such that a soverign chain basing off the settings here unchanged could allow for the same trust with the relay and other siblings. Issues happened in this light, addressed in #696 (at least this was attempted, not merged now I see 😬 )

[bkchr]

a parachain template ready to interact with a Relay chain and its Common Good parachains

I don't really get why we need a template for everything. These should be tutorials and not just random templates. In the end there would be thousands of templates, while they actually just have 2-3 xcm things configured differently.

[hbulgarini]

@PierreBesson is rococo-parachain used in the testnets?

[PierreBesson]

@hbulgarini No, we are not using rococo-parachain for any of our deployed testnets.

@bkchr Could rococo-parachain have been the historical source used to build the tick chainspec ? Maybe it would be nice to revamp it and make it accessible for users of the CLI with something like --chain staging-parachain to build "dumb parachains" chainspecs and replace the tick, trick and track raw chainspecs.

[bkchr]

Tick/Trick/Track are currently based on rococo-parachain.

of the CLI with something like --chain staging-parachain to build "dumb parachains" chainspecs and replace the tick, trick and track raw chainspecs.

Yeah that could be done.

[hbulgarini]

Tick/Trick/Track are currently based on rococo-parachain.

@bkchr are we still using Tick/Trick/Track? If we do, then we will keep parachain runtime. If not, i would say that we might remove it.

[PierreBesson]

Yes we are using Tick parachains for testing parachain scalability in our Versi internal network. We could replace them by a different runtime without much effort. However Tick parachains are very convenient for us because they don't need any injected aura keys to produce blocks.

[hbulgarini]

Ok, perfect. Based on this discussion rococo-parachain is still needed. Please be aware that as part of polkadot-parachains folder restructuring (#1240), this parachain will be moved to testing folder.