Positive and negative provenance in database systems
cse590q: Database Seminar (Winter 2010)
Wednesdays 2:30-3:20 in CSE 405 (new database lab!)
Please sign up for the class email list: 590q
In databases, provenance describes the source and derivation of data in tables and views. In scientific applications, provenance is typically regarded as important as the data itself, as it provides proof of its validity, and ensures repeatability of the result. Technically, provenance is related to topics like database annotations, information integration, probabilistic databases (599t/wi09), and consistency & trust in community databases (590q/au09). In this seminar, we will shortly review the three principal types of provenance (why, where and how), then focus on some broader aspects: dealing with nested subqueries, approximating provenance, explaining missing query answers ("negative provenance") and citing entries in online databases. The choice of topic is partly motivated by our current work on causes and explanations for query results.
One day before the day of your presentation, prepare and send out a one-slide graphical
summary of your paper; scanned hand drawings are fine and even encouraged (Examples).
For the last class, pair up with another student and prepare a one-slide graphical big picture
summary of the discussed papers (Examples). Present this slide at the beginning of the last class.
Addition: Along with the summary slide presenters are now required to send out focus points of the paper, for example: "Introduction, section 3.2, definition 2.1 and example of figure 5". All participants should at least read these focus points to prepare for the discussion.