Conveners
Andreas Athenodorou, The Cyprus InstituteEd Bennett, Swansea UniversityJulian Lenz, Swansea UniversityConveners
Embedded Files
General Motivation
General Motivation
Pillars of the Open Science according to UNESCO's 2021 Open Science recommendation.
For the results of computations to be meaningful, they must be reproducible—when anyone applies the same analytical techniques to the same data, they must get the same results. For other researchers to be able to verify that a result is reproducible, and build on the result and techniques used to obtain it, the data and code used must be shared openly.
In this unit you will learn ways to ensure that your results are reproducible, and how to apply the principles of open science to build on others’ work and help them build on yours.
Prerequisites
Prerequisites
Beyond the Essentials, in this unit we assume knowledge of a programming language, and will use examples specific to Python. If you don’t already know a programming language, then we would recommend working through Software Carpentry’s Introduction to Python before diving in.
Lectures
Lectures
Open Science
Open Science
Open Publications
Open Publications
Topics
Topics
Green and Gold open access
Creative Commons Licensing
The arXiv
Institutional policies and repositories
Rights retention
Links
Links
Open Lattice Data
Open Lattice Data
The International Lattice Data Grid
The International Lattice Data Grid
Links
Links
Open Software Development
Open Software Development
Topics
Topics
Version control
GitHub
Issue tracking
Pull requests
Automated testing
Agile techniques
Links
Links
Introductions to Git
Repository hosting platforms:
Open Workflows
Open Workflows
Topics
Topics
Software licensing
Repositories for open workflows
When to combine, or separate, data and code
Disposable vs reusable tools
Links
Links
Reproducibility
&
Open Science
Reproducibility
&
Open Science
Reproducibility
Reproducibility
Essentials of Reproducibility for LFT
Essentials of Reproducibility for LFT
Topics
Topics
Types of reproducibility
Precisely specifying hardware
Environment specification – containers and alternatives
Ensuring reproducibility with random numbers
Links
Links
Reproducible Parallelism for LFT
Reproducible Parallelism for LFT
Reproducible Configuration Generation
Reproducible Configuration Generation
Topics
Topics
Persisting RNG state
Recording provenance of field configurations
Field configuration and configuration metadata storage formats
Links
Links
Reproducible Measurements in LFT
Reproducible Measurements in LFT
Topics
Topics
Tracking provenance of measurements
Data formats for measurement inputs and outputs
Reproducible Data Analysis in LFT
Reproducible Data Analysis in LFT
Topics
Topics
Tracking provenance of analyses
Data formats for analysis inputs and outputs
Code structure
Workflow and data-flow management
Removing manual steps
Links
Links
Workflow managers
Libraries for statistics in lattice
Reproducible Data Presentation in LFT
Reproducible Data Presentation in LFT
Topics
Topics
Outputting tables
Styling plots
Keeping data consistent
Maintaining reproducibility when working with notebook-based programs
Removing manual steps
Links
Links
Library
Library
As this is a relatively young and rapidly-developing field, there are few textbooks but many online resources available.
Books:
Books:
Irving, et al.’s Research Software Engineering in Python (free online, also available in print) provides a good well-rounded introduction to many of the principles. Many chapters of this borrow from some of the lessons above.
Online lesson material:
Online lesson material:
You should also check with your institution’s library for guidance on your institution’s policies and procedures around open research.
Page updated
Google Sites
Report abuse