Hi all,
Currently, for a dataset processed with a given pipeline <pipeline>, the root directory derivative dataset is named <pipeline>/.
One of the issues that was discussed in the BIDS derivatives breakout group is how to label a derivatives directory to indicate specific run parameters, which would (for instance) allow comparing the results of the application based on the presence or absence of a given parameter. As concrete examples, consider iterating over registration methods, or smoothing windows.
The consensus proposal was to permit generic labels, so that derivative datasets may also be named <pipeline>_<label>, where label may be specified by the user or algorithmically constructed by the app itself.
I want to propose a `--exec-label <label>` command-line option that BIDS apps respect, to apply the given label to the derivative dataset name. I'm not strongly attached to the option name, but I do believe that some such access should be permitted.
What do y'all think?
--
Chris Markiewicz