Application Developer- Ab Initio

These skills are essential because the applications now exist as Ab Initio graphs rather than COBOL
programs.
Proficiency in the Ab Initio Graphical Development Environment (GDE): Building modifying
and debugging graphs using standard components (Reformat Join Sort Rollup Normalize
Lookup etc.) custom transforms and embedded code.
Understanding and Assessing Auto-Converted Graphs: Graphs produced by Ab Initios
automated COBOL/IMS conversion tool are not standard hand-built graphs. They follow
machine-generated patterns - often verbose deeply nested or structured in ways that differ
significantly from graphs written from this environment these converted graphs must
be assessed and modified to implement new requirements or fix defects. This requires the
ability to trace generated logic back to the original COBOL source identify the relevant transform
or component within an auto-generated structure and make targeted safe changes without
disrupting the surrounding converted logic.
Metadata Management: Working with the Enterprise Meta Environment (EME) for version
control dependency analysis impact analysis and data lineage.
Parameter Handling: Using Parameter Definition Language (PDL) effectively.
Orchestration and Workflow: Conduct It (or Express It) for scheduling and managing job flows
- this largely replaces JCL and IMS transaction management. Job scheduling is handled via
Atomic Automation which orchestrates Ab Initio workloads in the production environment. A
critical aspect of this environment is that Atomic Automation workflows contain parallel job
dependencies - multiple jobs may execute concurrently with interdependencies that must be
understood when diagnosing failures or assessing the impact of a change. This is distinct from the
sequential step-by-step flow within an individual job; the broader workflow topology must also be
considered.
Data Flow Traceability and File/Dataset Lineage: A critical problem-solving skill in this
environment is the ability to trace data content backwards and forwards through job flows
and workflows - following a file or dataset from its point of creation through each transformation
it undergoes across jobs graphs and workflow stages. This includes understanding what
populates a file how it is transformed at each step where it is consumed downstream and how
parallel workflow paths may contribute to or depend on its content. This traceability underpins
three core data concerns that must always be considered:
Data Integrity: Ensuring that transformations preserve the accuracy and consistency of data
values as they move through the system - detecting where values may be incorrectly
computed overwritten or corrupted relative to what the original IMS application would have
produced.
Missing Data: Identifying conditions under which records or fields may be absent dropped
or skipped - whether due to filtering logic join mismatches conditional branches or
upstream job failures - and understanding the downstream impact of that absence.
Data Retention: Understanding how long data persists at each stage - which files are
transient (used within a single run) which are retained across cycles and how GDG-style
generational patterns control the lifecycle of datasets. Knowing what data is available for
how long and under what conditions is essential for recovery reprocessing and audit
support.
Data Processing and Integration: Handling large-scale ETL/ELT processes including migrated
IMS segment data copybooks EBCDIC packed decimal and zoned decimal formats.
Administration and Operations: Co Operating System runtime management monitoring
logging error handling deployment and performance tuning (parallelism multifile systems
resource optimization).
Testing Validation and Move to Production (MTP): This is a multi-layered discipline in the
converted Ab Initio environment and must be treated as a structured process not an afterthought.
Unit Testing of Converted Graphs: Changes to auto-converted graphs require targeted unit
testing at the graph or component level - isolating the modified logic constructing or
sourcing representative input data and verifying that outputs match expected results relative
to the original IMS behavior. Because the converted code was machine-generated even
small changes can have non-obvious ripple effects within the surrounding graph structure;
unit testing must be thorough and deliberate.
Data-Driven Validation: Test cases must be grounded in real or representative data -
including edge cases common in the original IMS environment (e.g. packed decimal
boundary values missing segments GDG rollover conditions). Comparing Ab Initio output
against known-good baseline results from the original system (or a prior run) is the most
reliable validation approach.
End-to-End and Integration Testing: Because jobs within workflows have parallel
dependencies changes must be tested not just at the graph level but across the full job flow
- verifying that upstream outputs feed correctly into downstream jobs and that no parallel
branches are disrupted.
Move to Production (MTP) Coordination: MTP in this environment requires understanding
and coordinating multiple interdependent activities: packaging and promoting Ab Initio graph
changes through the EME; updating or validating Atomic Automation workflow definitions if
job dependencies change; confirming that MFS screen-related graph changes are consistent
with the deployed screen definitions; communicating the scope and timing of changes to
operations and business stakeholders; and verifying that production data files and GDG
generations are in the correct state prior to cutover. A practitioner must also understand the
rollback implications of a failed MTP - what state files and workflows will be in and what
steps are needed to recover.