Natural
Software AG's English-like fourth-generation language for building business applications on the mainframe, tightly integrated with the Adabas database and still maintained for enterprise use today.
Created by Software AG (principally Peter Pagé and Margit Neumann)
Natural is a proprietary fourth-generation programming language (4GL) created by Software AG and first launched in 1979. Designed for building data-intensive business applications, Natural pairs an English-like, statement-oriented syntax with deep, native integration to Software AG’s Adabas database. Its central promise was developer productivity: programmers could express common business operations – reading records, looping over result sets, formatting reports – in fewer lines than the contemporary third-generation languages such as COBOL or PL/I. Decades later, Natural remains in active use across banking, insurance, government, and other enterprises, and Software AG has publicly committed to supporting it well into the future.
History & Origins
Natural emerged from the same German software house that had already made its mark with Adabas, the high-performance adaptable database system Software AG released in 1971. As Adabas adoption grew through the 1970s, customers needed a faster, higher-level way to write applications against it than assembler or COBOL with database call interfaces. Software AG’s answer, introduced in 1979, was Natural – a language whose development is generally attributed principally to Peter Pagé and Margit Neumann, with the work reportedly taking shape over the preceding years from around the mid-1970s.
The design goal was to make application development approachable by basing the syntax on natural-language-like English keywords, hence the product’s name. A short Natural program could open a database loop, select and filter records, and produce a formatted report with only a handful of statements. Because Natural and Adabas were engineered together, the language could exploit Adabas features – such as descriptor (key) fields and inverted lists – directly and efficiently.
Design Philosophy
Natural was built around a few enduring ideas:
- Ease of use first. The language deliberately favors readable, declarative-feeling statements over low-level mechanics. Database access, screen handling, and reporting are first-class language constructs rather than library calls.
- Tight database integration. Natural is at its most natural with Adabas: statements like
READ,FIND, andHISTOGRAMmap onto Adabas access methods. Over time, support was added for other data sources, but the Adabas pairing remains the canonical one. - Productivity for business logic. The language and its companion tools (the Predict data dictionary, Natural Construct for code generation, and others) aimed to let small teams build and maintain large transaction-processing systems.
- Longevity and stability. Natural targets organizations that expect their core systems to run for decades, so backward compatibility and predictable behavior have been treated as priorities.
Key Features
Reporting Mode and Structured Mode
Natural supports two programming styles. The original reporting mode is terse and report-oriented: loops and conditional blocks can be delimited loosely (for example with DO / DOEND), and a single END can close constructs. As applications grew more complex, Software AG introduced and recommended structured mode, which requires explicit loop-closing statements (such as END-READ for a READ loop and END-ENDDATA for an end-of-data block) and does not let a bare END close processing loops. Structured mode yields programs that are easier to read and maintain, which is why it became the preferred style for new development.
A simple reporting-mode example reads employee records by key and prints two fields:
| |
Database Statements
Natural’s data-access verbs are part of the language itself:
| Statement | Purpose |
|---|---|
READ | Read records in physical, ISN, or descriptor (key) sequence, initiating a processing loop |
FIND | Select records matching search criteria on descriptor fields |
HISTOGRAM | Read the distinct values of a descriptor and their counts |
STORE / UPDATE / DELETE | Modify the database within transaction logic |
END TRANSACTION / BACKOUT TRANSACTION | Commit or roll back work |
Screens, Reports, and Maps
Beyond data access, Natural provides built-in facilities for interactive screen handling (often through maps, reusable screen definitions) and for producing formatted reports with statements such as DISPLAY, WRITE, and control breaks like AT END OF DATA and AT BREAK.
Strong, Static Typing
Natural variables are declared with explicit data formats and lengths (for example alphanumeric, numeric, packed, date, and time formats), giving the language strong, static typing suited to financial and record-keeping work where precise numeric handling matters.
Evolution
Natural began life on the IBM mainframe but expanded over the years. Software AG ported it to open systems under the Natural for Linux, UNIX and Windows (LUW) line, broadening the database options it could work with – on the mainframe alongside Adabas it can target data in VSAM, Db2, and IMS/DB, and on LUW it can work with Adabas, Oracle, Db2, and other relational systems.
Tooling modernized as well. For many years Natural development happened in mainframe editors and Software AG’s own environments; in 2009 the company released NaturalONE, an Eclipse-based IDE that brought source management, debugging, and integration features to Natural projects and made it easier to expose existing programs as services for web and cloud front ends. NaturalONE has continued to advance, reaching the 9.2 generation by early 2024, and Software AG has published an Adabas & Natural Community Edition to let developers evaluate the stack.
Crucially for its customers, Software AG announced in 2016 that Adabas and Natural would be supported through 2050 and beyond – an unusually explicit long-term commitment that reflects how deeply these systems are embedded in their owners’ operations.
Current Relevance
Natural occupies a particular niche: it is rarely chosen for greenfield startups, but it underpins an enormous amount of running enterprise software. Banks, insurers, government agencies, telecoms, and manufacturers still operate mission-critical Natural applications, many of them decades old and still processing high transaction volumes reliably. As a result, much of today’s Natural activity centers on maintenance and modernization – integrating existing programs with modern interfaces, refactoring into structured mode, exposing business logic as APIs, and selectively migrating where it makes sense, all while preserving proven code.
The product’s ownership has also shifted. In 2024, following Silver Lake’s acquisition of Software AG, the Adabas & Natural line was organized to operate as a standalone business, signaling continued, focused investment in the platform’s user base.
Why It Matters
Natural is a clear example of the fourth-generation language movement that sought to raise the abstraction level of business programming above COBOL and PL/I. By treating database access, screen handling, and reporting as built-in language constructs, it let comparatively small teams build and sustain large transaction-processing systems. Its longevity is itself instructive: Natural shows how a well-fitted language-and-database pairing can remain economically valuable for decades, and why “legacy” systems are often kept running – and modernized in place – rather than rewritten. For anyone studying the history of enterprise computing, Natural and Adabas together tell an important story about productivity, durability, and the long tail of mainframe software.
Timeline
Notable Uses & Legacy
Banks and Insurance Companies
Many large financial institutions built core banking, policy administration, and claims systems in Natural over Adabas during the 1980s and 1990s. A substantial number of these systems remain in production, valued for their throughput and reliability on the mainframe.
Government and Public Sector
Tax authorities, social-security agencies, and other government bodies have used Natural and Adabas to run high-volume record-keeping and citizen-service applications, where stable, long-lived systems are essential.
Telecommunications and Utilities
Telecom operators and utility companies adopted Natural for billing, provisioning, and customer-management systems that needed to process large transaction volumes against Adabas databases.
Manufacturing and Logistics
Enterprises in manufacturing and distribution used Natural to build inventory, order-processing, and ERP-style applications, taking advantage of its rapid development model for data-intensive business logic.
Application Modernization Projects
Because so much business logic still lives in Natural, the language features heavily in modernization efforts -- exposing existing Natural programs as services, integrating with web and cloud front ends through NaturalONE, or migrating selectively while preserving proven code.