Principles of Provenance

Recent research in a variety of settings (databases and data warehouses, geographic information systems, scientific workflows, grid computing, and the Semantic Web) has addressed the problem of keeping track of metadata about creation and modification history, influences, ownership, and other provenance or lineage information. Such metadata is essential for making informed judgments about data quality, integrity, and authenticity. In addition, ideas about provenance are now being used in several areas of computer science such as probabilistic databases, operating systems, file synchronization, and annotation propagation. Other topics, such as version control and archiving, may also benefit from better understanding of provenance. We believe the time is ripe to develop the foundations of the topic and address questions such as:

• What does it mean for information to be "provenance"? What is and what isn't provenance?
• What kinds of problems does provenance address, and how does one characterize correct solutions?
• How does one compare different models of provenance?
• Why is provenance so hard to get right, even though it seems rather obvious?
• Where should research efforts be focused to make the best progress?