Is there a reason we use the

<container_platform>.registry

in the config and not include it with the module itself? https://github.com/nf-core/tools/blob/99b23f0366232d793a82e71f28b2e621c382b889/nf_core/pipeline-template/nextflow.config#L245-L249 It seems to cause problems for users outside of nf-core who want to use nf-core modules and don't know this also needs to be included into their config ( or are not using a slimmed down nf-core template )

my guess is that this makes it easier to switch around, in case we switch it. looks like @Adam Talbot added it initially https://github.com/nf-core/tools/pull/2233

> my guess is that this makes it easier to switch around This. At the time, we had 2 big things moving: • AWS had published a mirror for biocontainers (but didn't include mulled 🙄) • quay.io fell over every 5 minutes and any users who need reliability needed to secure their supply chain and have a private mirror. The default docker behaviour can use a docker.registry but allow for a manual override, so setting docker.registry is the best(ish) compromise, unless you want to write a 3000 line config file for every pipeline update. Recently, this is less important because the AWS mirror has gone and we have tools such as Wave to use for mirroring.