Hands-on Reproducible and Scalable Brain Imaging Analysis with Nipype note 1
joeyh at
Docker is not super well suited for use in reproducing scientific research (command line complexity; need to GC old sessions etc), but they are using it anyway. The singularity container system may be better suited, won't be surprising if scientific communities switch to it.
Christopher Allan Webber likes this.
Docker is also too insecure for use in HPC cluster environments.
Christopher Allan Webber likes this.