Go through the typical workflow and "relate" to reproman workflow

[yarikoptic]

Go through https://github.com/ReproNim/containers#a-typical-workflow
and relate to reproman AWS EC2 cluster workflow: https://github.com/ReproNim/reproman#step-2-create-analysis-datalad-dataset-and-run-computation-on-aws-hpc2 .

Long shot/back reference: running bids-apps using reproman https://github.com/ReproNim/reproman#a-typical-workflow-for-reproman-run

[yarikoptic]

mriqc could be the realistic/useful workflow to use so we do not need to wait for fake-fmriprep to emerge.
Can take any of the https://datasets.datalad.org/?dir=/openneuro (aka ///openneuro) datasets. Note that mriqc outputs already produced using reproman and stored in https://github.com/OpenNeuroDerivatives/ , so could be even fun to see how results would differ. Might be nice to provide https://diffoscope.org/ reports between them.

[yarikoptic]

if can't login to discovery with reproman, you can ssh directly, install reproman there in some conda environment and use local orchestrator

[yarikoptic]

May be @jbwexler could give a version of that example script we have in README.md on ttps://github.com/ReproNim/ds000003-demo but how he would setup/run on TACC (slurm, shared filesystem).