The Discovery Search Engine

1
The Discovery Search Engine

• Peter Buch

2
Mission

• Offer the most precise, easiest to use XML and
  SGML search engine on the market.
• Build partnerships with the leading information
  providers.
• Promote the eXtensible Retrieval Language as the
  standard for the semantic web.

3
Meaning over mechanics

• Index and search your collection the way people
  understand it.
• Meaning belongs on the outside, mechanics on the
  inside
• XRL maps logical retrieval views to physical
  elements of your database

4
Two kinds of maps

• Some markups add meaning
• Title, author, abstract, quotation
• These make up the Semantic Map
• Some markups add structure
• Section, chapter, paragraph, body
• These make up the Structural Map
• (Some dont do anything!)

5
The XRL mapping syntax

• Uses RDF-like XML serialization syntax
• Semantic map uses pattern matching to equate
  constructs to Plain-English labels
• Structural map defines things you can rank,
  retrieve and combine

Sample XRL schema
6
The search engine

• Sensitive to XRL at both indexing and retrieval
  time
• Supports multiple retrieval views
• Supports XML and SGML
• Will join multiple DTDs and databases, local and
  remote
• Builds composite (précis) documents

7
Advantages

• Adapts to users language and needs
• Logical indexing reduces index size
• Joins DTDs and architectures in a single,
  consistent logical interface
• Scoring, ranking and retrieving of any size
  snippet
• Build and save composite documents

8
Our competition

• Some are query driven
• Treat retrieval like SQL
• Some are structure driven
• Accidental indexing
• We are unique in providing an adaptive logical
  layer between database and search engine

9
Our customers

• Providers, owners, custodians of collections of
  high-value XML and SGML documents
• Consolidators of scientific, legal, medical and
  scientific databases
• Technical documentation clearinghouses
• Government agencies, universities
• ASP and VSP

10
Where we are

• Working prototype completed
• Detail plans for production release well under
  way
• Incorporated
• Patents application in progress

11
For more information

• Read more about XRL
• XRLFAQ.htm
• Call me
• 919-932-1224
• Email me
• peter_buch_at_hotmail.com
• Coming soon
• www.xmldiscovery.com