Oninit® BLT — Recover What the 4GL Knows Before the People Who Wrote It Are All Gone

The Informix 4GL application that runs the business today was written by a team that mostly retired five years ago. The documentation, if it exists, is fifteen years out of date. The new team is reading source code and guessing. Every project that depends on the application — a cloud migration, a regulatory report, a customer-facing rewrite — stalls on the same question: what does this application actually do? Hiring around the gap is slow. Replacing the application is a multi-year program.

Oninit® BLT gives you a third option: extract the application's behaviour in machine-readable form — workflows, fields, rules, database operations, messages, transactions — with evidence attached to every claim, so the next system can be designed against known facts rather than reconstructed from interviews.

The business case

• Modernisation that starts from facts. The spec is the input to whatever comes next — a code rewrite, an LLM-driven regeneration in a modern stack, a re-platforming project. The new team works from evidence-cited behaviour, not from reading the original source by hand.
• Audit-grade output. Every workflow, rule, and database operation is cited to one or more underlying facts. The output is deterministic — identical inputs produce byte-identical results — so the spec is safe to lodge as a record and re-run as the application changes.
• LLM-ready handoff. The streamlined JSON projection is designed to feed directly into a code-writing LLM. The path from "we have a legacy 4GL application" to "we have it regenerated in a modern stack" starts with BLT's output.
• Lower dependency on tribal knowledge. The spec is durable. It does not retire, change jobs, or forget. When a project depends on knowing what the application does, the answer is a file on disk rather than a calendar invite with a developer who's been there twenty years.
• Optionality on platform choice. The behavioural recovery is independent of the target. The same spec supports a rewrite in Python, Java, .NET, or a cloud-native stack — the decision is the organisation's, not a constraint inherited from the extraction step.

Why Oninit

Oninit has supported the global Informix community for two decades as “the Down System Specialists.” BLT is built and supported by the same team that recovers other people's Informix outages — the people who know the database engine, the 4GL runtime, and the compiled binary format in the most demanding conditions production ever produces.

What it removes from your risk register

• “We don't know what the 4GL application actually does any more.”
• “The original developers have all retired and the new team is guessing.”
• “We can't move off Informix because we'd have to rewrite an application nobody fully understands.”
• “The regulator is asking for documentation we don't have.”

Next step

A first proof-of-value runs in days, not months — point BLT at the application, the schema, and a compiled binary, and you have a sample spec ready for the modernisation team to read inside the first session. The Worked Examples show the kind of output a single application produces; the next step is running it against yours.