Search-based application: Difference between revisions

Revision as of 14:27, 2 September 2010

Revision as of 00:50, 5 November 2010

Line 3:

== Pre-Conditions ==

Search based applications are fully packaged applications that:

SBAs are typically deployed when there is a need to synthesize heterogeneous content from multiple sources (text documents, multimedia files, database content, etc.), or, more commonly, they are used to replace traditional [[Relational database|relational database systems]] as the primary data access infrastructure when one or more of the following constraints exist:

 Are built on a search backbone to enable sub-second access to information in multiple formats and from multiple sources

 Are delivered as a unified work environment to support a specific task or workflow, for example: eDiscovery, financial services regulatory compliance, fraud detection, voice of the customer, sales prospecting, pharmaceutical research, anti-terrorism intelligence, or customer support.

 Integrate all the tools that are commonly needed for that a specific task or workflow, including:

 Multi-source information access

 Authoring

 Collaboration

 Business process

 Reporting and analysis

 Alerting

 Visualization

 Provide pre-configured data integration with multiple repositories of information in multiple formats as appropriate for the application ___domain.

 Integrate ___domain knowledge to support the particular task, including industry taxonomies and vocabularies, internal processes, workflow for the task, connectors to specialized collections of information, and decision heuristics typical of the field.

 Provide a compelling user interface and interaction design that eliminates the need for users to “pogo stick” or continually jump from one application to another. This buffers the user from the complexity of operating separate applications and enables them to focus on getting work done.

 Are quick to deploy, easy to customize or extend, and economical to administer

Source: Worldwide Search and Discovery 2009 Vendor Shares and Update on Market Trends, IDC #223926, July, 2010 by Susan Feldman and Hadley Reynolds.

* ''' High volumes of data and/or users:''' <br />Search engines are optimized for fast processing of access requests ([[I/O|read operations]]), while databases are optimized for recording and storing transactions ([[I/O|write operations]]). Accordingly, search engines can provide efficient processing against very large data sets by a high number of simultaneous users (whether human users or other software applications).

* ''' A need for real time information:''' <br />To maintain the referential integrity of data and balance transactional and access requirements, large databases are typically updated in [[Batch processing|batch processes]], resulting in [[Latency (engineering)|data latency]]. Search engines maintain indexes using continual incremental and differential update processes, supporting real-time data availability.

* ''' A need for ad hoc access and/or reporting against a broad range of criteria:''' <br />Search engines enable users to access and manipulate data according to any criteria maintained in the index— whether the criteria is extracted from a source system or created by the engine during the course of natural language processing—without advance programming of all queries and views.

* '''The need to extend access to specialist systems to untrained users:''' <br />Because SBAs can extract data from complex systems and make it available in an independent data layer that can be accessed using technologies such as natural language search, [[fuzzy matching]] and [[Faceted classification|faceted navigation]], SBAs are being used to extend access to the content in complex, specialist systems to non-specialists.

== Practical Uses ==