[tracking] Pain points about chitin HOT 2 OPEN

• Downloading a dataset from a remote (could be achieved with a v.similar mechanism to uploading)
• Running a dataset with different operational parameters (need a way to distinguish between runs of the same dataset, also need a way to generate new UUID but still tie to parent UUID)
• Ensuring the integrity of state between local and remote (especially if chitin not yet on remote?)
• Running results/saving graphics - would be cool if this was totally automated
  • Munging scripts could report to chitin (with a UUID, and whatever parameters: exactly like creating a dataset)
  • Quick and dirty version could copy (or otherwise instruct) graphics to be saved in a certain place
  • Could automatically generate plots as part of the check_integrity (could even expand this: eg. FASTQ could always have fastqc reports generated if they are missing)

from chitin.

• Destroying a dataset

from chitin.