Glossary

A glossary of common terms used in the TRACE project.

Reproducibility

Obtaining consistent results using the same input data, computational steps, methods, and code, and conditions of analysis [16].

Transparency

The reporting of materials and methods in a manner that provides enough information for others to independently assess and/or reproduce findings (adapted from FASEB [29]). Research object (RO): Standards-based method of bundling and describing research artifacts, typically via linked-data technologies [31].

Computational artifacts

Inputs, code, outputs, documentation and relevant details of the computational environment required to ensure the transparency and reproducibility of results obtained through computational means.

Input

any file, database, or stream that is used by a computational workflow (or workflow step) to obtain reported results. Examples include data and configuration files.

Code

Scripts and/or source code that are not part of an installed software package and used to execute steps of a computational workflow used to obtain reported results. This includes code for data preparation, analysis, modeling, plotting, etc., and is considered separate from the operating system or any installed software packages (see environment below).

Output

Information created by a computational workflow (or workflow step) used to obtain reported results. Examples of outputs include: cleaned or otherwise pre-processed datafiles used for analysis; analytical or simulation results; tables, figures, and numerical results; log files and other program outputs, messages, and errors (e.g., stderr, stdout, log files).

Environment

The specific set of operating system and dependent software versions used to execute a computational workflow used to obtain reported results. The computational environment may be described as a list of operating system and software versions or as a virtual machine or container image.

Documentation

Materials providing additional information required to reproduce reported results including step-by-step instructions for obtaining or accessing data; setting up or accessing the environment; and running the complete computational workflow.

TRACE workflow

The complete sequence of steps taken to create and release a TRO from author-provided artifacts. TRACE workflows may be manual or automated and may include actions such as disclosure avoidance measures.

Sensitive data

Information that may be regulated by law due to possible risk of harm due to their disclosure.

Proprietary data

Information for which the rights of ownership are restricted so that the ability to freely distribute the data is limited. (USGS)

Transient data

Information that is not persisted.

Ephemeral data

Information that change rapidly over time and may be lost if not collected immediately.

Streaming data

Information that is generated continuously, generally in large volumes, and must be processed sequentially and incrementally over time (adapted from https://aws.amazon.com/streaming-data/). Examples include information from social networks, financial trading floors, and telemetry from connected devices.

Disclosure avoidance

Methods of protecting confidentiality that may require modification of data. (https://www.census.gov/topics/research/disclosure-avoidance.html)