In general, you need enough documentation for a naive user to completely replicate both experiments and every analysis in the paper. This is needed both for reviewers and/or scientists who want to follow up on our work, and also for you when, in a few months, we all forget the specifics of what we've done and want to either revise our manuscript or follow up on the experiments. We want to make it clear what code to run and what each step does. In general the entire repo needs some more documentation, and careful review for consistency and usability. This repository provides a nice model to follow.
For example:
[x] in data_analysis_code documentation and notebooks, rename experiment 1 to "Sustained Attention Experiment" and experiment 2 to "Variable Attention Experiment"
[x] add a README file in data_analysis_code describing which notebooks do what
[x] in all notebooks, change paths to be relative paths, assuming the jupyter server is started in the directory containing the given notebook. and add a note that you're assuming the jupyter server is started in the directory containing the given notebook so that users know how to run the notebooks.
[x] in the stimulus_generation_code folder add a readme (a) crediting Megan deBettencourt for providing stimuli and preprocessing code and (b) describing what should be done with the zipped files. also specify that the .m scripts need to be run in MATLAB (specify minimum version number if known and/or the version number you used)
[x] each experiment needs a readme file either describing the experiment or pointing to the relevant sections of our paper (with hyperlinks to the preprint). also describe in the readme how each experiment's directory is organized and how to run the experiments. it also looks like running the experiments requires the stimuli-- so you should explicitly link to the zip files and describe what needs to be done to run the experiments.
[x] in each experiment's data folder, you should describe (completely) what the different files are, how the data in each file are formatted, etc. also link to code for parsing (processing, analyzing) the data.
In general, you need enough documentation for a naive user to completely replicate both experiments and every analysis in the paper. This is needed both for reviewers and/or scientists who want to follow up on our work, and also for you when, in a few months, we all forget the specifics of what we've done and want to either revise our manuscript or follow up on the experiments. We want to make it clear what code to run and what each step does. In general the entire repo needs some more documentation, and careful review for consistency and usability. This repository provides a nice model to follow.
For example:
data_analysis_codedocumentation and notebooks, rename experiment 1 to "Sustained Attention Experiment" and experiment 2 to "Variable Attention Experiment"data_analysis_codedescribing which notebooks do whatstimulus_generation_codefolder add a readme (a) crediting Megan deBettencourt for providing stimuli and preprocessing code and (b) describing what should be done with the zipped files. also specify that the .m scripts need to be run in MATLAB (specify minimum version number if known and/or the version number you used)