INFO Features



 Overview

  • Extends the INFO Productivity Language to include free text Information Retrieval.
  • Allows the creation of databases and complete applications which require a mix of structured and free text data items.
  • Gives users the ability to enter, update, interrogate and report on both types of data with the same command language.
  • Preserves the tabular view of Relational databases while allowing any fields to contain unlimited amounts of formated text.
  • This introduction is designed to provide a description of the additional facilities which INFO-Text provides to the INFO user and should be read as a sequel to the introduction to INFO.
  • More detailed information can be obtained from the INFO Reference Manual which describes the formats, functions and options available to each of INFO-Text's commands.



 Background

The 1970s saw the development of two fundamentally different approaches to data management which can be described as structured and free text. The structured methods are powerful for dealing with highly complex relationships between data elements and are appropriate for use in bill of material applications and many others. Free Text systems, on the other hand, were designed to tackle legal and, later on, library automation tasks where the data has a less rigid structure. The free text systems provide access to the stored information through an inverted list of pointers to all significant 'words' in the text body.

Like the more structured DBMS products, free text systems support query languages for data access but the differences are tremendous. While DBMS queries are particularly strong on value searches (i.e. FIND UNPAID INVOICES OVER 1,000 AND INVOICE-DATE GREATER THAN 1/1/96), free text query languages can efficiently support word, phrase, synonym and proximity searches on even large amounts of more 'natural language' text data (i.e. FIND MICROPROCESSOR WITHIN THE SAME PARAGRAPH AS 'AERONAUTICAL NAVIGATION').

Unfortunately, many important applications require the strengths of both DBMS and free text systems to be successful and it has thus been a dream of those who have used both to see them integrated into a single environment.

INFO-Text brings the two together and enables applications to store, retrieve, update and report on any mix of structured and free text data with the same command language.


 INFO-Text's Extensions to INFO

INFO-Text offers all of the facilities of INFO, a well established productivity language for relational data management, personal computing and Rapid Application Development. In addition, INFO-Text delivers new power in the following features:

Existing File Compatability

INFO-Text extends the INFO tradition of working with existing files to include any text in ASCII form created by external editors or word processors. This feature makes possible a new range of important applications such as correspondence tracking systems, contracts administration and the linking of application source code with specifications and documentation.

Concordance

INFO-Text provides a full range of user-controlled concordance options which enable INFO's query commands (RESELECT and ASELECT) to also locate occurrences of words, parts of words, phrases and synonyms. Proximity searches with Boolean conditions and wild cards are easily performed.

Dictionary

INFO-Text maintains a complete dictionary of concorded words for each database which allows control of stopwords and is used to test and document search strategies.

Full Screen Editor

INFO-Text includes a VDU independent, full screen editor which is automatically invoked whenever text items are addressed with the ADD, UPDATE or INPUT FORM commands.

Mixed Retrieval

INFO-Text's retrieval commands (LIST, DISPLAY and PRINT) have been extended to support any mix of fixed width and free text data with new controls for text column widths, tab positions and line or page spacing. As with INFO, data can be retrieved from a selected and up to 9 RELATEd files at any time.

Relational Model

The data model preserves the tabular user view and flexibility of Relational data management while allowing any attributes or columns to contain free text.

The Data Model

INFO-Text databases consist of any number of flat tables or files which can be dynamically joined through common data items with the RELATE command.

Free text data is held separately and accessed through pointers which are either generated by INFO-Text itself or are specified by the user. It is this second option which enables INFO-Text to access text data which has been created externally. As the individual records need to contain only the pointer information to the variable length text they are physically stored in a fixed width format which has been defined by the user.