rdahis / paper_template Goto Github PK

Flow for R&R revisions.

1. Centralize all numbered comments in one document uploaded to /extra/submissions/'entry'/comments.txt
2. Assign comments to issues with clear tasks described in text. Assign issues to people, add tags, add milestone for "journal resubmission", etc. Add issue numbers below each review comment so that it can be tracked 1:1.
3. Work on issues, add sentences/paragraphs together with commit messages describing changes made to code and writing.
4. As issues start to be closed via PRs, put together a 'release' text, which will feed into letter to the editor and referees.
5. When all is done, set a tag for the release and send it off for the journal.

Tags: review, build, analysis, writing, negative replies.

See:

• https://github.com/gslab-econ/template
• https://www.ripl.org/wp-content/uploads/2018/08/RIIPL_data_infrastructure_20180613_preprint.pdf

See https://github.com/hpellegrina/hello-world-econ/blob/master/documentation/github-manual.pdf

• Set up GitHub with versioned folders /input, /output, /code.
• Set up Dropbox with non-versioned folders /extra, /tmp, etc.
• Set up local folder with symbolic links to Dropbox folder.

Create script for automatic setup of symbolic links with local paths and user name.

Described here: https://www.overleaf.com/learn/how-to/Dropbox_Synchronization

Flow seems to be simple:

1. Set up Overleaf-Dropbox integration for folder /products/paper.

Any edits between Dropbox and Overleaf get synced.

Important: the "true" state is always on Github. Overleaf is useful for simultaneous edits and comments, but things always have to be committed on Github.

Feature

• Motivation
• Desired enhancement

Bug

• Expected behavior
• Actual behavior
• Environment

Large files in /extra, /input, /output, /tmp.

Ideal requirements:

• Having data versioning.
• Allowing for "releases" to be completely reproducible versions of what was done at the time.

Options:

• Add to .gitignore and leave files only in Dropbox.
• Git LFS

Turn it into:

• Clean datasets according to different levels of observation
• Merge & covariates for various types of analysis.

Generate variables and other stuff inside analysis dofiles if specific to that analysis.

• https://style.tidyverse.org/

build
analysis
products
enhancement
bug