What are the main differences (strengths and weaknesses) between checkpoint and other reproducibility packages like packrat?

Checkpoint is a more comprehensive package from a reproduction standpoint. This is because it works in tandem with a server side solution, the checkpoint-server. Most of the burden is on the checkpoint-server, not the end user. Packrat does not have a concept of client or server, and a user could still find themselves in "dependency hell". Packrat requires a user to rsync, scp, or some how get your Packrat file and code to another user. That can be a cumbersome and tedious process.With checkpoint package, all you need to reference in your code is the snapshot date, and then you can commit your R code to a public or private repository, such as github.com. Then all another user has to do is pull your R code down, run it, and they will be using the same packages as you, because they are using the same snapshot date.In short, the checkpoint-server brings order to the world of R packages. It knows exactly how things looked at a given point in time, and everyone sees the same stuff, therefore everyone is on the same playing field.

For more information please review this page:
http://projects.revolutionanalytics.com/documents/rrt/rrtpkgs/
Note This is a "FAST PUBLISH" article created directly from within the Microsoft support organization. The information contained herein is provided as-is in response to emerging issues. As a result of the speed in making it available, the materials may include typographical errors and may be revised at any time without notice. See Terms of Use for other considerations.
Properties

Article ID: 3104102 - Last Review: 10/31/2015 13:08:00 - Revision: 1.0

Revolution Analytics

  • KB3104102
Feedback