Is there a reason why it would be "better" in any ...
# nostupidquestions
b
Is there a reason why it would be "better" in any way to have a Nextflow workflow (being run in Seqera cloud, if that matters) use a Wave container (which doesn't already exist, I'd have to make it) rather than a pre-existing Docker container?
p
It's a thing where there might be differing opinions. I feel a wave container is better as it can't suddenly randomly go away. In case wave containers are immutable by default (without using a checksum tag) that's a bonus. Another bonus is that a wave container is at least partially a known entity - one can see what went into making it. Apart from those aspects, it's about the same. There might even be reasons it's better not to use custom built ones (e.g. restricted environment require opening to more places, preference for "standard" images, et.c.)
thankyou 1
m
I think, the important question here is, what you are referring to with a Nextflow workflow. If it is a workflow that you build only for yourself and your lab, then using a pre-existing Docker container is an absolutely viable option, even sound according to the Keep it simple principle. Even more so, if you have build it yourself and have access to the Dockerfile. But if the workflow is meant to be shared, the extra effort for better reproducibility pays off on the long term - say somebody needs a Singularity container or needs to run on a different architecture. Wave will handle that for you and the nf-core tooling will be tested for that case as well.
gratitude thank you 1