Computational provenance

TY - JOUR

T1 - Computational provenance

AU - Silva, Claudio T.

AU - Tohline, Joel E.

N1 - Funding Information: Systematic mechanisms for capturing this viding users with meaningful interpretations of information are at the heart of a new field of process executions. These interpretations aim to research called computational provenance. Most dic-explain provenance in a way that’s closer to how tionaries define provenance as an object’s source domain experts see it. or origin—a record of an item’s ultimate deriva- tion and passage through various owners. Ulti- mately, provenance helps determine an object’s e hope this theme issue raises value, accuracy, and authorship. But in addition awareness and contributes to a to enabling result reproducibility,1 provenance for better understanding of the is-computational tasks and the data they manipulate W sues surrounding computational and derive has other benefits as well. In particular, provenance to the broader CiSE community. Reit helps users interpret and understand results—in search and technology being developed in this some cases, it can be more important than the ac-area have the potential to be transformative and tual results themselves. improve how people do science in a variety of do-The articles in this special issue show many uses mains. By examining the sequence of steps an ex-of provenance that go beyond result reproducibil-pert followed to produce a result, a user can gain ity. The first article—“Provenance for Computa-insights into the chain of reasoning used, learn by tional Tasks: A Survey,” by Juliana Freire and her example, and potentially reduce the time to insight. colleagues—targets potential users of provenance Combined with social networking, provenance technology who aren’t quite experts on the topic. could even serve as a catalyst to mass collabora-It covers the key issues involved in capturing, stor-tion. Likewise, given the ease with which we can ing, and querying computational task provenance share digital information, the provenance infra-and describes some existing systems. structure currently available could serve as strong The next two articles discuss ongoing research motivation for authors to publish, along with their projects and their use of provenance informa-scientific articles, data, and codes, the actual pro-tion. In “Provenance in High-Energy Physics cess they used to solve a problem. Provenance is Workflows,” Andrew Dolgert and his colleagues destined to gain a higher profile in coming years as describe their efforts as part of a large-scale in-the broad field of computational sciences matures ternational collaboration of physicists analyzing and more strongly emphasizes the reproducibility data from CERN’s Large Hadron Collider. Their of archival simulation results. research involves petabytes of data accessed by thousands of systems and collaborators, so they’re Acknowledgments aiming for a software-based infrastructure that This work was funded by the US National Science will replace the traditional lab notebook. In “Prov-Foundation, the US Department of Energy, and an enance in Comparative Analysis: A Study in Cos-IBM Faculty Award. mology,” Erik Anderson and his colleagues focus on a collaborative visualization framework to help analyze data from the Cosmic Code Comparison Project, which aims to establish the robustness of results from a set of cosmological simulations. The large number of simulations, plots, graphs, and visualizations involved has rendered manual Cláudio T. Silva is an associate professor at the Uni- bookkeeping of result provenance nearly impos-versity of Utah. His full vita information appears on sible. In particular, the authors describe how they p. 21. Contact him at csilva@cs.utah.edu. customized the VisTrails system to get the type of high-end visualization required. Joel E. Tohline is an alumni professor at Louisiana The last two articles discuss specific aspects of State University. He has a PhD in astronomy from the provenance management technology. In “Prov-University of California, Santa Cruz. Contact him at enance: The Bridge between Experiments and tohline@lsu.edu. Data,” Simon Miles and his colleagues present an innovativeuseofaprovenanceprojectcoupledwith Editor’s Note: Although Cláudio T. Silva is listed as a workflow engine. Finally, in “Problem-Solving a guest editor and helped pull together the articles Methods for Understanding Process Executions,” for this special issue, he wasn’t involved in the peer- Jose Manual Gómez-Pérez and Oscar Corcho ap-review process at all (co-guest editor Joel E. Tohline proach provenance from the perspective of pro-handled this duty).

PY - 2008/5

Y1 - 2008/5

KW - Computational provenance

KW - Data provenance

KW - Reproducibility of computation results

KW - Scientific data management

UR - http://www.scopus.com/inward/record.url?scp=42449138798&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=42449138798&partnerID=8YFLogxK

U2 - 10.1109/MCSE.2008.71

DO - 10.1109/MCSE.2008.71

M3 - Editorial

AN - SCOPUS:42449138798

SN - 1521-9615

VL - 10

SP - 9

EP - 10

JO - Computing in Science and Engineering

JF - Computing in Science and Engineering

IS - 3

M1 - 4488059

ER -