Skip to content

my comments and Satra answers to the lessons 1-3 (copy of the email conversation) #10

Description

@djarecka

Lesson 1: Core concept
General:
- you're saying at the beginning that the lesson use the Simple Workflow, but it's not clear to me how specific parts are related to the repo. I would expect much more guidance.

can you suggest what kind of guidance?

Element 1:
- are you planning to "convert" the JSON file (that is missing for now) to the two standards you mention? It would me very useful IMO.

we should revise this. NIDM-E is not set in stone, so if we release we should have some confirmation of a NIDM-E version.

Element 2:
- don't understand the first sentence, "...when a different dataset containing the same data or a slightly different workflow is used."

i don't see this sentence here: http://www.reproducibleimaging.org/module-dataprocessing/02-concepts/

Lesson 2: Annotate..

General:

  • there are no "elements"

elements were conceptual pieces in the concepts lesson - we can refer back to those elements, but don't expect elements to be in everything (i think).

Links don't work:

http://data.wu.ac.at/csvengine/csvm
the second looks like a markdown typo.

Data and Metadata

  • not sure what is the main point here, convincing people to submit data to archives? A few sentences of intro might be nice. When do you use NDA and when NeuroVault for sharing?

yes, and introductory statement would be good there.

Lesson3: Create and...
General:
- there are no "elements"
- IMO a general motivation when to use VM, when Docker or Singularity, in scientific application, is missing. You have it in your presentation, but not here.
- as I mentioned, it might be useful to work on the final version of the videos so they are easier to use.

we should update things here with respect to the presentation.

Docker
- I know you mention Bids-App, but it could be useful to point to some Python or Nipype images, so they don't have to install everything to run Nipype workflow.

sure - the current nipype dockerfile has most things necessary.

Singularity

  • I don't see any "smooth transition" to Vagrant. I understand that Vagrant is needed when one uses OSX or when has to create an image on a machine without sudo (since she/he is a root in VM), but it's not obvious IMO.

we should point out a few more things here, but there are a few places in the web we can draw from.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions