Reproducibility and Open Science Catalog
This is a draft page for my website to list various open source or open access projects I am involved in.
[Add some general info about ROS and pointers to guidelines, etc.]
Where to find my work:
ORCID 0000-0003-1384-4504 This unique identifier is useful for identifying distinct authors. It is a good idea to create one for yourself if you don't already have one. Click this link to find citations of some papers and grant proposals.
Semantic Scholar profile This is a new search tool for academic papers that seems promising.
Currently it only includes computer science publications so many of my papers don't appear.
Preprints on arXiv. Many authors post preprints on this open access preprint server.
On my website. Most publishers allow posting some version of a paper on the authors website.
Some pdf files of papers are available at this link, in some cases linked to our institutional repository UW ResearchWorks so that they are permenently available.
Code and data to accompany publications:
For recent papers my group have strived to make the computer code and any data used in producing figures or tables for the paper available to readers, whenever possible in an open access repository with a permanent DOI and a licence attached so that others know how they can be reused and how they should be cited.
Recently we have been using Zenodo, which also links to GitHub so that "releasing" a version of code there automatically triggers a new version on Zenodo, with a zip file of that commit of the code and a new DOI.
Here are some examples:
We also sometimes just create a GitHub or Bitbucket repository, e.g.
Version control and code-sharing repositories:
I primarily use Git for version control, not only for code being developed but also for writing papers, class notes, grant proposals, etc.
While some things are in private repositories, many things are public and can be browsed from the links below.
Open source software projects:
Clawpack is an open source project I started in 1994 and is now actively developed by a team of researchers scattered at many institutions. We use this GitHub organization to work together effectively.
GeoClaw is a part of Clawpack and is the part of the code I am most actively working on. Originally developed for tsunami modeling, it is used for other geophysical flows such as storm surge.
We use Pull Requests to review code submitted to this software, which
facilitates having a discussion about proposed changes.
Each of the Clawpack repositories has its own set of pull requests, which can be found on the GitHub pages.
Here are a few examples of discussions I have been involved in recently:
- Matplotlib Parallel Plotting Capabilities
- GoogleEarth plotting
- bc2amr calling sequence wrong in amrclaw routines
- Faster bound
Activities on Reproducibility and Open Science:
I have been active in promoting reproducibility and open science, for example as part of the UW eScience Institute working group on this topic.
Slides from some talks can be found among the seminars.
Some other activities:
Producing a Journal Article on Probabilistic Tsunami Hazard Assessment, a short case study for this project
Setting the Default to Reproducible: Reproducibility in Computational and Experimental Mathematics Developed collaboratively by participants in an ICERM workshop I helped organize in 2012.
Top Ten Reasons to Not Share Your Code (and why you should anyway), appeard in SIAM News, 2013.
Reproducible Research for Scientific Computing: Tools and Strategies for Changing the Culture by R.J. LeVeque, I. Mitchell, and V. Stodden, reporting on another workshop from 2011.