Contextual Search Using Bond Graphs in Collaborative Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search mechanisms in collaborative environments are inefficient as they rely on brute force methods, lacking the ability to consider relationships between diverse types of searchable items and context, leading to irrelevant results.
Innovation Solution
A search mechanism that captures and utilizes relationships between searchable items, represented as 'bonds', to improve search efficiency by forming a connected graph where nodes are items and edges are bonds, allowing for targeted and ranked searches based on logical connections and user-defined criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If brute force search methods are used to search through all items in a collaborative environment, then the search can find all matching items, but the search efficiency is low and the results are not ranked by relevance
Solution Approach 1:
The system pre-establishes bond relationships between items before search operations occur. These bonds represent various types of associations (authorship, reference, collaboration, etc.) and are stored in advance, allowing the search to quickly traverse meaningful connections rather than examining all items systematically.
Solution Approach 2:
The patent introduces 'bonds' as intermediary elements that connect search items. These bonds serve as mediators that encode relationship information between items, allowing the search mechanism to navigate through the collaborative environment via meaningful connections rather than direct item-to-item comparison.
2Productivity
If conventional search mechanisms search only within file system directories, then the search scope is limited and fast, but the search cannot access people and context information
Solution Approach 1:
The search mechanism is designed to operate across multiple types of entities uniformly. The same bond-based traversal works for searching documents, people, meetings, and other collaborative artifacts, allowing a single search system to handle diverse query types and return relevant results from across the entire collaborative environment.
Solution Approach 2:
The patent adds a new dimension to the search space by incorporating relationship-based connections beyond the traditional file system hierarchy. Instead of only navigating through directory trees, the search can traverse bond relationships that connect items across different organizational dimensions (authorship, reference, collaboration, etc.).
3Device complexity
If all searchable items are treated homogenously in a search, then the search mechanism is simple, but the search cannot distinguish between different types of items and their relationships
Solution Approach 1:
Different types of bonds represent different qualities of relationships between items. The system distinguishes between various bond types (authorship bonds, reference bonds, collaboration bonds, etc.) and uses the appropriate bond type based on the search context, allowing the search to traverse relationships with the correct semantic meaning for each situation.
Solution Approach 2:
The search mechanism changes parameters based on the bond types encountered during traversal. Different bond types can have different weights, traversal priorities, or filtering criteria, allowing the system to adjust the search behavior dynamically based on the nature of the relationships being traversed.
Data Source
AI summary
Techniques are provided for using bonds, which reflect relationships between items, to facilitate searches against the items. The degree of separation between any two given searchable items is based on the minimum number of bonds that have to be traversed to arrive at one of the two searchable items when starting at the other of the two searchable items. The bonds are used to respond to a search request, performing a search relative to a particular searchable item. For example, the search may involve only those searchable items that are within a particular degree of separation of the particular searchable item. As another example, the search may involve determining the order in which searchable items are compared against search criteria of the search request based on the degree of separation of the searchable items from the particular searchable item.