TRUST 2014 - 1st International Workshop on Reproducible Research Methodologies and New Publication Models

It becomes excessively challenging or even impossible to capture, share and accurately reproduce experimental results in computer engineering for fair and trustable evaluation and future improvement due to ever rising complexity of the design, analysis and optimization of computer systems, increasing number of ad-hoc tools, interfaces and techniques, lack of a common experimental methodology, and lack of simple and unified mechanisms, tools and repositories to preserve and exchange knowledge apart from numerous publications where reproducibility is often not even considered. After focusing on systematic and community-driven program and architecture auto-tuning and co-design combined with machine learning and crowdsourcing during past 6 years we faced numerous, practical challenges related to reproducible experimentation. Based on this experience and feedback from the community, we decided to organize this workshop as an interdisciplinary forum for academic and industrial researchers, practitioners and developers in computer engineering to discuss challenges, ideas, experience, trustable and reproducible research methodologies, practical techniques, tools and repositories to:
capture, preserve, formalize, systematize, exchange and improve knowledge and experimental results including negative ones
describe and catalog whole experimental setups with all related material including algorithms, benchmarks, codelets, datasets, tools, models and any other artifact
validate and verify experimental results by the community
develop common research interfaces for existing or new tools
develop common experimental frameworks and repositories
share rare hardware and computational resources for experimental validation
deal with variability and rising amount of experimental data using statistical analysis, data mining, predictive modeling and other techniques
implement previously published experimental scenarios (auto-tuning, run-time adaptation) using common infrastructure
implement open access to publications and data (particularly discussing intellectual property IP and legal issues)
enable interactive articles
Submission guidelines
Easychair submission website is open.
We invite papers in three categories (please use these prefixes for your submission title):
T1: Extended abstracts should be at most 3 pages long (excluding bibliography). We welcome preliminary and exploratory work, presentation of related tools and repositories in development, experience reports, and wild & crazy ideas.
T2: Full papers should be at most 6 pages long (excluding bibliography). Papers in this category are expected to have relatively mature content.
T3: Papers validating and sharing past research on design and optimization of computer systems published in relevant conferences. These papers should be at most 6 pages long (excluding bibliography).
Submissions should be in PDF formatted with double column/single-spacing using 10pt fonts and printable on US letter or A4 sized paper. All papers will be peer-reviewed. Accepted papers can be published online on the conference website that will not prevent later publication of extended papers. We currently arrange proceedings to be published in the ACM Digital Library.
Important dates
Abstract submission: March 7, 2014 (Anywhere on Earth)
Paper submission: March 14, 2014 (Anywhere on Earth)
Notification: April 14, 2014
Final version: May 2, 2014
Workshop: June 12, 2014
Workshop organizers
Grigori Fursin, INRIA, France (Collective Mind / cTuning project)
Bruce Childers, Alex K.Jones and Daniel Mosse, University of Pittsburgh, USA (OCCAM project)