Where exactly should we store the presentation code
We started discussing where exactly should we store the presentation code.
There are two options and both seem to have pros and cons.
-
In the
code
directory within the dataset. We currently recommend this approach. However, BIDS is not explicit that this is the place for presentation code.Pros:
- Presentation code resides with the data.
Cons:
- The data repository has to be cloned to work on presentation code.
- The histories of data and presentation code interleave.
- The data repository must be cloned to execute the presentation code.
-
As a separate repository. This is consistent with BIDS
Code
key in the Stimulus presentation details.Pros:
- The presentation of the code evolves separately from the data.
- Different versions of the code can be tagged and used for acquiring different datasets.
- Two datasets can be acquired with the same code without duplicating it.
- The presentation code can be tested and deployed without cloning the data repository.
Cons:
- The presentation code resides outside of the data. The correct version must be linked.
Please contribute in the comments.
Edited by Mateusz Pawlik