Convera Technologies International Ltd
integrates with: RetrievalWare is designed for search electronic information across an
enormous range of distributed sources.
RetrievalWare is a secure, highly scalable
platform for mission critical search and categorization in the
enterprise. Built on robust, field-proven technology, RetrievalWare’s distributed architecture provides high
performance infrastructure to index, search, categorize and link
information across an enormous range of distributed sources. The
platform’s cartridge architecture allows the system’s
functionality to be tailored to best address industry specific
problems, language requirements, knowledge discovery or other
information management needs.
Optimized Search Precision & Recall
Through the combination of advanced search
methods and unique categorization & classification technologies,
RetrievalWare achieves optimized precision & recall across vast
and diverse information sources. Offering three search methods –
concept, pattern and Boolean – which can be used independently
or interactively, users have the flexibility to search the way
they want with confidence in the results.
Concept search does what we naturally do in
conversation with each other – account for the individual
differences in the way we express similar ideas. Using Convera’s
unique and powerful Semantic Networks, RetrievalWare can search
for concepts or units of meaning instead of merely simple terms
and exact word matches. With millions of concepts and terms
interlinked by millions of word relationships, the Semantic
Networks help to pinpoint all relevant meanings without
requiring users to define every possible way of expressing their
Concept search is a powerful tool, addressing
major challenges in delivering high precision and recall.
However, the form or spelling of words may vary widely as the
result of misspellings in original documents or subsequent input
errors which may occur during OCR and data entry, or as the
result of legitimate alternative spellings. In these instances
Convera’s patented Adaptive Pattern Recognition Processing
(APRP™) uses a sophisticated vote and rate scheme that considers
a number of different features of the pattern instead of just
character pairs. It can find words even when the word is
misspelled by the searcher, within the document text or where
there are legitimate spelling differences.
RetrievalWare also provides a powerful
Boolean or keyword search mode, which is a very efficient,
flexible and accurate information retrieval tool. This form of
search is also useful as a filter when used in concert with
RetrievalWare’s concept and/or pattern search to limit and rank
results. RetrievalWare’s powerful indexing and search engine
uses advanced linguistic analysis to ensure high precision and
recall for both boolean and concept search modes.
Categorization and Dynamic Classification
Users can launch a search and automatically
classify the results based on pre-defined or dynamically
generated classifications. The underlying taxonomies can consist
of Convera’s pre-packaged industry taxonomies, customer defined
taxonomies or custom taxonomies supplied by Convera’s Taxonomy
Development Partners. By combining these classifications, users
are able to isolate and discover critically relevant information
that other search and retrieval software cannot uncover.
Categorization and Dynamic Classification
represent a behavioral and technological leap for users and
organizations alike. Rather than being forced to fit searches
within the constraints of inflexible categories, users can
dynamically create their own information categories based on the
context of their search at the moment. Further, those categories
can inter-relate and display information from widely disparate
sources and locations, permitting users to discover knowledge
that might have otherwise remained hidden.
Dynamic knowledge discovery helps
organizations better share information and expertise, make
faster more informed decisions and reap the full value of all
the enterprise’s human and information assets.
Repository and Data Access
RetrievalWare’s secure, distributed
repository access is managed by Convera’s Synchronizers, which
support more than 200 documents types found within file servers,
groupware systems, relational databases, document management
systems and web servers. The Synchronizers automatically
recognize modifications, changes or updates to information
assets system-wide and automatically update the indexes to
maintain accuracy of each information repository.
Convera Family of Synchronizers
Complete and accurate search results can only
be achieved when your information management solution has access
to all the necessary content available. Convera RetrievalWare
offers a family of synchronizers for establishing access to the
most common repositories where enterprise information is stored.
With RetrievalWare's Synchronizers you can index and retrieve
documents from a variety of repositories with just a few simple
RetrievalWare Synchronizers automatically
perform three key functions - locating documents, notification
of document changes and extending access controls of the
repository. The synchronizer’s primary function is to locate all
the assets in a given repository and deliver them to the
RetrievalWare indexing engine. Then the Synchronizer watches the
repository and detects when existing documents are changed,
removed or when new documents are added, so the index can be
updated. RetrievalWare’s Synchronizers not only handle top-level
documents in a repository, they also handle their attachments.
This is particularly important for Groupware repositories such
as Lotus Notes and MS Exchange.
Synchronizers work with RetrievalWare’s Real
Time Agent Profiles option to give users the ability to set up
custom profiles. Profiles enable the users to identify the types
of documents for which they should receive automatic
With the help of the Synchronizers,
RetrievalWare provides a security infrastructure that extends to
and leverages the native security of the various document
repositories. RetrievalWare Synchronizers use the access control
list of the repository to determine the access rights of each
document. This access information is then included in the index
so RetrievalWare delivers only the documents for which the user
has access rights. To maintain appropriate security, the
Synchronizer monitors each repository’s access control list for
permission changes and passes them to RetrievalWare.
Through a modular approach, the RetrievalWare
architecture allows any combination of Synchronizers to be used.
The standard RetrievalWare offering includes a file system
Synchronizer. Additional Synchronizers are available for
relational database management systems such as MS SQL, Oracle,
Sybase, Informix and Teradata; document management systems such
as Documentum and FileNET; and groupware servers such as
Microsoft Exchange and Lotus Notes. If an organization’s
solution requires access to a repository or file type not
currently supported, the RetrievalWare Developer Extensions SDK
includes the AFM Toolkit that allows customers or Convera’s
professional services group to build the necessary
Scalability is more than just fast searching.
Rather, it is the ability to maintain search performance even
when the demands on the system rise by orders of magnitude.
RetrievalWare is designed from the ground up for the highly
robust, secure environment required in enterprise-class mission
critical applications. RetrievalWare adapts to deliver maximum
performance on every network regardless of size, load, or
configuration. Furthermore, the implementation of Java 2
Enterprise Edition (J2EE) enhances the RetrievalWare platform
with a modern, flexible architecture, employing industry
standards to achieve broad integration support with leading
development and deployment environments, including Web Services
and Microsoft .Net. New, pure Java APIs enable the rapid
deployment of new features and functionality as well as easier
integration for customers, partners and OEMs.
Scalability and Modularity
Because search and classification are used
both within and outside enterprises with a variable number of
users and dynamic content, it is very important that the
enabling technology be able to grow to support changing needs
with a minimal effect on performance. RetrievalWare was
developed from its inception on a distributed process
architecture, which provides the foundation for building
scalable solutions that can maintain performance even as demands
on the system rise by orders of magnitude. RetrievalWare can
leverage additional server machines and maintain search
performance over large databases and large numbers of users.
RetrievalWare provides an integrated security
infrastructure that extends and leverages the native security of
modern information management and document management
repositories that are connected to the RetrievalWare knowledge
space. RetrievalWare’s crossrepository, cross-platform security
protocol ensures that users only gain access to the documents
and libraries they are authorized to see, preserving security of
the total system. For categorization & classification, the
administration of folders is done via a secure login so that
there is control over who can create and edit the categories.