What happened
A new feature called bitemporal provenance has been added to TypeGraph, an open-source library designed for managing graphs in SQL databases. This feature allows users to track facts in two different ways: valid time, which shows when a fact was true in the real world, and recorded time, which indicates when the database captured that fact. This adds a layer of complexity and power, especially when dealing with derived facts that depend on multiple sources.
Why this matters
The introduction of bitemporal provenance changes how databases handle facts that may have uncertain origins. For example, if a vulnerability is identified based on conflicting sources, retracting one source doesn't simply remove the fact; it re-evaluates which other facts still stand. This capability enhances data integrity and reliability in decision-making processes, particularly in fields like cybersecurity, where accurate information is critical.
Context
Typically, databases lack the ability to manage time-based data effectively. While some systems have implemented temporal tables, many require extensions to support system-versioning. TypeGraph, however, offers a unique solution by explicitly storing historical data and allowing users to perform point-in-time reads without the need for database-level support. This approach is inspired by existing truth-maintenance systems which help track the justification of facts over time.
What this means
With bitemporal provenance, users can confidently navigate complex data environments where facts can change over time. By computing which facts still have support after a retraction, the system prevents misinformation from cascading through dependent data. This not only facilitates better audit trails but also empowers users to make informed decisions based on the most accurate and up-to-date information available, paving the way for more robust data management practices in various industries.
