Indexing Quoted Text in Conversations for Unified Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional email messaging systems fail to support efficient conversation-based searching due to the prohibitive cost of maintaining data structures for both message-based and conversation-based searching, limiting users' ability to perform simple conversation-based searches.
Innovation Solution
A system and method that index quoted text in messages within conversations, allowing for advanced conversation-based search capabilities by distinguishing between original and quoted text, and enabling users to switch between conversation-based and message-based searching without duplicating data structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate data structures are maintained for message-based searching and conversation-based searching, then both searching capabilities are enabled, but the storage cost and processing complexity increase prohibitively
Solution Approach 1:
The patent segments the indexing process by creating separate index components: original text index components for direct message terms and quoted text index components for terms from quoted passages. This segmentation allows the system to maintain a single unified index structure that serves both message-based and conversation-based searching, avoiding the need for separate data structures while preserving both search capabilities
Solution Approach 2:
The unified index structure serves multiple functions simultaneously. The same index that stores original text components also stores quoted text components, enabling the system to perform both message-based searching (using original text components) and conversation-based searching (using quoted text components) without requiring separate dedicated indexes for each function
2Adaptability or versatility
If quoted text is indexed with original text, then conversation-based searching is enabled, but the index size and processing overhead increase
Solution Approach 1:
The patent applies local quality by differentiating between original text and quoted text through separate index components. This allows the system to mark and distinguish quoted passages while maintaining the overall index structure. The local differentiation enables conversation-based search functionality without requiring a complete duplication of the entire index, thus managing data volume growth
3Ease of operation
If messages are organized into conversations, then contextual understanding is improved, but the ability to search individual messages efficiently is compromised
Solution Approach 1:
The patent segments the indexing process by creating separate index components: original text index components for direct message terms and quoted text index components for terms from quoted passages. This segmentation allows the system to maintain a single unified index structure that serves both message-based and conversation-based searching, avoiding the need for separate data structures while preserving both search capabilities
Solution Approach 2:
The system performs preliminary indexing of both original text and quoted text components when messages are received or updated. This preliminary action ensures that when search queries are executed, the necessary index components are already in place and ready for immediate retrieval, maintaining search efficiency without requiring real-time processing delays
Data Source
AI summary
A method for indexing quoted text includes, obtaining a plurality of index entries in an index. A respective index entry corresponds to a respective term in one or more messages, and has a plurality of index components; a respective index component identifies a message associated with the respective term. The method further includes, responsive to receiving a first message; updating the index by: associating the first message with a conversation having one or more other messages; identifying text in the first message as quoted text based on text that occurs in one or more of the other messages; and updating, in the index, first-message index components that each includes an identifier of the first message, which includes original text index components and quoted text index components. The quoted text index components include information indicating that the corresponding terms are terms in the quoted text of the first message.