The Desktop Search Challenge, Part 1

I have been doing lots of testing and thinking on desktop search lately. Since my thinking is constantly evolving, I decided to snapshot some thoughts in my blog as new ideas emerge.

Essential Requirements
1. Quality Search Results
- integrate and correlate search results from all my data sources
- rich file format support including pdf, mp3, photos, zip &c.
- rank results first by date, than by "topic"
- quick preview of a search result

2. User Experience
- multi-modal user interface (more later...)
- UI and visualization to help me discover hidden data relationships
- gentle on system resources
- easily accessible privacy controls
- no "unwanted" ads

3. Data Mobility
- federated search across my computers and web archives
- indentify duplicate documents and deduce document version
- optional sychronization (index or content) with cloud-based storage

4. Outlook Integration
- thread e-mail messages; needs to be more robust than gmail (I don't understand why Outlook still stores "sent mail" into its own separate folder.)
- a threaded message becomes the "document," which provides better context than a individual e-mail
- a learning thesaurus for nick names and abbreviations (e.g. M$ == microsoft)
- select which mail folders to index...Google Desktop Search (GDS) doesn't offer this option today
- index contacts, to-dos, and calendar items

5. Privacy
- no personal data get transferred beyond my computer except during an authorized, federated P2P search
- authenticated user; respect the application or file systems' security settings in search results
- deleteing a file or message triggers the removal of the file from the index
- permit user to "blacklist" certain file, folder, and file types from the index
- select which file folders to index...GDS indexes the entire disk today
- verified identity and strong encryptions for P2P communications and data transfer