The goal of the workshop on Failed Aspirations in Database Systems (FADS 2017) is to create a forum for discussing systems ideas that seemed promising but ended up failing in practice, along with revisiting prior systems work with an eye toward validating ideas that were once thought bad (or vice versa) through in-depth analysis or further experimentation. Results in the database systems field focus almost exclusively on novelty and research successes. However, as system builders, researchers, and practitioners, lessons from failures are often just as important as the successes. We learn hard lessons when implementing systems. For example, ideas that seemed good on paper don’t work (or work poorly) when implemented in non-trivial software systems, or ideas require significant refinement in non-obvious ways after initial implementation. In addition, systems ideas and techniques once thought of as “failures” might be ripe for re-evaluation (or vice versa).
This workshop provides a venue for disseminating negative/strange/unexpected results, ideas, and “war stories” with the goal of helping future researchers and practitioners avoid the hard lessons learned by others. FADS will be organized as a one-time workshop as a way to try something new, not as a yearly recurring event.
The scope of the workshop will encompass topics such as:
- Explanations and analysis of why an idea that seemed good failed after system implementation
- In-depth analysis or experimental results justifying why a once failed system idea is promising
- Studies that refute earlier findings with proper justification, or explain cases where a “good” earlier system idea or technique performs poorly
- Explanation or analysis of why an idea from academia did not pan out in a commercial system