How can you transform without documentation?
Well, it does exist, the 100% reliable program documentation: it is the hexadecimal object code that Njema uses as the ultimate information.
Only the object code determines the logic and the flow of a program. So if you are able to use it, the generated Cobol program will always be the exact logical image of the assembler program.
Njema therefore requires as input all fully expanded conversion lists generated with HLASM or ASSEMBH.
Why Cobol as a target language?
Cobol was developed at a time when mainframe programs could only be written in autocoder and assembler. In order to increase the productivity of programmers, Cobol was developed in such a way that the assembler programmers quickly found their way around the new language.
As a result, Cobol is a kind of“assembler in disguise“. In its structure and mode of operation, it is very close to the assembler thought pattern. It is therefore technically very obvious to go to Cobol as a first step. From there, you can easily move on to other languages just as automatically.
Using a demo program, we show how Cobol and two Java variants derived from it are automatically created from an Assembler program.
Thanks to the integrated refactoring algorithms, every Cobol generated is well structured and easy to read. This means that it can be handed over immediately to regular program maintenance.
Methodology
However, it is essential for the success of the project to know all the details of every single line of application code in advance.
- Therefore, the Njema Assessment examines all called system and user macros, coding conventions, as well as tricks and habits of the programmers.
- It analyzes register usage, detects self-modifying code and stores everything in the Njema Metadata Database.
- Njema also creates a complete inventory of all data fields, their properties and their relationship to each other.
- Since the metadata of all entities is available when a single object is transformed, many rules, e.g. for re-engineering, can work with respect to all other statements of all other entities.
- Based on the metadata, a project-specific transformation engine is built by selecting hundreds of transformation rules from the Njema Rule Repository.
- This engine transforms all assembler programs in a single run.
- In each run, new possibilities for increasing the degree of automation are identified and then added to the transformation engine .
- Situations that occur rarely or only once are handled by singular, project-specific rules.
- After the last run, thetransformation engine has learned how to convert the entire assembler application into functionally identical Cobol: 100% automatically.
Monitoring
Large assembler projects in particular need to be carefully monitored and controlled. The transformation engine is therefore integrated into a monitoring process.
Fully automated assembler transformation
Complete automation has a number of advantages, it …Complete automation has a number of advantages, it …
- brings calm to the project
- dramatically reduces the risk of errors
- allows a good prediction of duration, effort and costs
- significantly shortens the project duration
- makes the project independent of regular program maintenance
- eliminates human error, which is otherwise unavoidable in large projects
- allows the transformation to be repeated as often as required without effort
- enables a complete transformation and the final integration test to be carried out within a few hours before the application can go into production as a Cobol incarnation.
Which platforms are supported?
The Njema Assembler Transformation is available for z/OS, TPF, VSE and BS2000.