Summary

1
Summary

• Contains most of the standard IR components
• DB as the central storage
• Data-driven
• Plug Play Data representation and
  Access/Manipulation tools
• Parallel
• XML for definition/rule files, RoR for 1st build,
  multiple programming language through directory
  services

2
Overview
Back-End
Crawler
Processed Files
Parser
Analyzer
Searchable Data
Searcher
Analyzed Request
Front-End
ReqAnalyzer
Request
Result
UI
3
Crawler
Communicate
Retrieve
Brokers
Def/Rule File
Schedule
Add
Add
Gatherers
Retrieve
Topic/Genre/Type Specific, PnP, parallel, Lead
broker exists
4
Parser
Retrieve
Parsing Controller
Def/Rule File
Schedule
Update
Generate
Parser
Input/Output-Format specific, PnP, parallel
Varying formats as needed by the analyzer
5
Analyzer
Retrieve
Analyzer Controller
Def/Rule File
Check
Schedule
Update
Generate
Analyzer
Input/Output-Format specific, PnP, parallel,
store result/pointer to Database
Data structures which are not in Database, e.g.
Inverted Index,
6
Searcher
Retrieve
Generate Use
Search
Return
Searcher
Searchable Data
UI
Def/Rule File
7
Request Analyzer
Retrieve
Use
Request Analyzer
Use
Generate
Analyzed Request
Knowledge Source
8
User Interface
Communicate
Search UI
Generate
Request
Browse UI
Communicate
Auxiliary Component
Retrieve
Result
9
Implementation Plan

• Stage 1
• Format/Rule/Data structure definition
• Communication framework
• Basic version of each component
• Toy corpus
• Stage 2
• Focus on back end
• Gather substantial corpus

10
Implementation Plan

• Stage 3
• Focus on front end
• System evaluation and analysis