Progress Through December 1999
Kernel Implementation
- Haystack is a simple operating system
- Separate privileged operation (persistence, security, etc.) from data-based services (field extraction, indexing, etc.)
- Integrate persistent, transaction-safe object store
Initial Exploration of Adaptation and Clustering
- Query history should influence future query behavior
- We integrated query infrastructure and improved UI
- Ready to begin application of learning techniques
- Explored use of clustering to organize Haystack data
- Preliminary theoretical work, prototyping