Kinetech · MARCUS — Managed Agent Resource Constructing Unbelievable Software

Your OpenEdge app runs the business. Your ABL bench is one resignation deep.

Progress ABL applications encode decades of how your company actually operates — in the data dictionary, the AppBuilder screens, and procedures that read like the business talking to itself. Marcus, Kinetech's AI Mendix developer, reads that source directly and rebuilds each application as modern Mendix: gated by your approvals, proven with published evidence, delivered to your own repository.

01 · The clock

The runtime is maintained. The situation isn't.

Let's be accurate: Progress still ships OpenEdge — 12.8 LTS is current and 13.0 arrived in February 2026. The problem is everything around your application: the 11.x line you may still be on, the ISV that built your app, and the people who can change it.

Apr 2025
OpenEdge 11.7 officially retired, per Progress's retirement notice — the whole 11.x base now runs without security patches
Oct 2028
end of the 12.2 LTS support window — the next forced re-platform decision, per Progress's published lifecycle
1-person
ABL "team" at many end-customer shops — a bus factor the business never signed up for
Scarce
"a growing shortage of Progress developers" — the ecosystem's own words, not ours

02 · The differentiator

Marcus reads your ABL directly

OpenEdge is one of the most machine-readable legacy platforms ever shipped. The data dictionary exports to a clean text format; AppBuilder screens are plain-text source; ABL itself is famously English-like — FOR EACH customer WHERE… reads like a specification. Marcus derives each application's full functional spec from the source, without workshops asking users to remember how a 2003 screen works.

OpenEdge artifactWhat Marcus extractsWhere it lands in Mendix
Data dictionary (.df export)Tables, fields, formats, indexes, validation expressions — plus field labels and help textDomain model: entities, attributes, associations — nearly 1:1
AppBuilder screens (.w)Frames, widgets, layout, and triggers — plain-text UI definitionsPages built on the extracted layout intent
Procedures & classes (.p / .i / .cls)Business rules, queries, calculations, batch routinesMicroflows and scheduled events
DB triggers & validationWrite/delete rules, referential logic, validation messagesValidation microflows and event handlers
Menus & navigationScreen flow and role-gated entry pointsNavigation model and page access
Users & permissionsApplication-level roles and gating as implemented in codeUser roles and entity access rules, verified by automated per-role tests

Why this matters: conversion tools transpile ABL into other languages; assessment consultancies inventory it. What hasn't existed is a developer that reads the source and then does the rebuild — tested per role, documented, and delivered as standard Mendix in your repository. That's Marcus.

03 · How a migration runs

Assess first. Rebuild what the business runs on. Prove every step.

The same four approval gates that govern every Marcus engagement are held by you:

Gate

Scope approval

The extracted specification of each application — what it does, who uses it, what carries forward — approved by you before a line of the model is written.

Gate

Permission ratification

The translated security model, presented in plain English for your sign-off — often the first time the app's real permission rules have ever been written down.

Gate

Design sign-off

Modern, branded design directions presented before pages are generated. Your users finally leave the gray ChUI/GUI screens behind — you pick what they land on.

Gate

Release acceptance

Each app reaches you only after all quality gates pass; it cuts over only when you accept it, evidence in hand.

04 · The date

Send the ABL source. Forty-five days to a running Mendix app for a typical OpenEdge module.

Here is the commitment we will put in writing. The assessment does not end with a range — it ends with a date, sized from your actual extracted complexity. Miss that date and the assessment fee comes back. We can promise a date because we never quote a calendar before reading your source.

The assessment names the date. Miss it, and the assessment fee comes back.

That is a real term, not a slide. It is affordable to us for one reason: the clock below starts only when a complete artifact pack is in our hands, and it pauses only for your own approval gates.

How the date gets sized

One OpenEdge functional moduleComplete artifact pack → UAT-ready
Compact — ≤50 procedures, one schema area, ≤2 integrations3–4 weeks
Departmental — 50–200 procedures, ABL classes, 2–5 integrations~45 days typical
Load-bearing — 200+ procedures, ISV-derived vertical, heavy AppServer or OERealm, multi-database10–16 weeks
Each further module in the same estate30–40% faster

Sizing model, not a price list — the assessment reads your actual source and returns one date. Bands assume a single rebuild boundary agreed at the scope gate.

What starts the clock

The date is only honest if the inputs are complete, so here is the whole list up front. Nothing on it is unusual, and most of it exists already.

What pauses it

Your gates, at your pace

Three approval gates are yours: scope, permission ratification, design sign-off. We answer each within three business days. Time beyond that on your side moves the date by exactly that much — and nothing else does.

What is outside it

Things we don't control

Production cutover on your change calendar. SSO and network provisioning by your IT. Third-party licenses or vendor contracts we don't control. Remediation of source-data quality problems the assessment flagged. Anything added after the scope gate.

The finish line

What "UAT-ready" means

Functional parity against the approved specification, running on your migrated data in a parallel environment, with production security enforced and tested per role, all nine quality gates green, and documentation published. Your users click through it.

What actually moves a date

ABL is highly readable and extracts well. OpenEdge estates take longer than that readability suggests for a different reason: they are usually large, and usually an ISV vertical with a local customization layer on top. Separating your customizations from stock product behaviour is the first real work, and it is exactly what the assessment is for. Temp-tables and ProDataSets used as a de-facto middle tier also need a design decision rather than a translation.

05 · Why Mendix

Keep the processes. Retire the dependency.

Your OpenEdge app survived this long because it fits your business better than any packaged alternative. That's exactly why the goal isn't a new ERP — it's the same processes on a platform mainstream developers can extend.

06 · Straight talk

What transfers, what changes, what we won't pretend

AreaThe honest position
Keyboard-driven UXHeads-down OpenEdge users live on function keys and muscle memory. We run a UX-parity workshop per app — keyboard shortcuts, inline grid editing, focused task pages — and we'll say it plainly: power users will be slower for a few weeks. Occasional users, new hires, and anyone on a browser or phone get faster on day one.
Server-side logicHeavy database triggers and batch ABL don't have to move on day one. During transition they can keep running server-side while Mendix reads and writes the same data — migrating them is a scheduled decision, not a leap.
ISV-owned codeIf your app came from a vendor, your right to the source depends on your license and any escrow. We check this first in the assessment — before anyone falls in love with a plan.
Reports & printOperational reports rebuild as pages and exports; pixel-specific print documents are a scoped workstream with our in-process PDF generation — stated per app in the assessment, never hand-waved.
Data qualityDecades-old databases carry surprises. Marcus profiles the data as it actually exists and surfaces exceptions as decisions in the spec you approve — mapped, defaulted, or archived — before migration.
Not everything should migrateSome modules are better served by off-the-shelf software in 2026. The assessment says so where it's true — you pay to rebuild what differentiates you, not what a $50/month tool does better.

07 · Quality & security

Every rebuilt app ships with evidence

Watch real build evidence land on the live Build Console, and see the delivery lifecycle at our interactive explainer.

08 · Engagement & pricing

Start with the assessment. Scale by the app.

Step 1 · Assessment

$6,000–$12,000

Source and dictionary export, automated read of your estate: a scored inventory (rebuild / retire / replace-with-SaaS), complexity and risk flags (trigger density, ISV source rights, data quality), and fixed-bid quotes for the first rebuild wave.

Credited toward onboarding if you proceed.

Step 2 · Delivered rebuilds

Fixed bid per app

Each app quoted from its actual extracted complexity — not a day-rate guess. Built, tested per role, documented, and delivered through the approval gates with published evidence.

Best for a first wave of 1–5 modules.

Estate scale · Most popular

Dedicated Marcus instance

For a full ERP-scale estate: a Marcus instance works your migration backlog continuously. Priced on what the work is, how many lanes run in parallel, and how fast you need it — always with Kinetech engineer review and an escalation SLA.

Bring your own Claude access: roughly 15–25% off the monthly.

Every figure here is an indicative range; your actual range is confirmed in writing before anything is charged. On bundled plans, heavy build months may incur metered usage beyond the included allowance; your quote states the allowance plainly. Your applications, models, and repositories are contractually yours.

09 · The questions you should ask

Fair pushback, straight answers

Our ERP vendor says stay on OpenEdge — they have a roadmap.

Then the assessment is cheap insurance either way. If your vendor is investing and your license covers what you need, the scored inventory tells you which satellite apps and customizations around the core are still worth rebuilding. If the "roadmap" turns out to be maintenance releases, you'll be eighteen months ahead of the shops that waited to find out. Either way you decide from evidence, not a renewal call.

The app works fine. Our ABL developer isn't retiring for years.

The bus-factor math doesn't wait for retirement parties. One person holding the only working knowledge of a system the business runs on is a risk your auditors would flag anywhere else. Migrating deliberately — while that person is available to answer questions and validate behavior — is the cheap version. The expensive version starts the week after their notice.

Why not an automated ABL-to-Java conversion? Some tools claim 100% automation.

They can convert it — that's real. The question is who maintains what comes out. The ecosystem's own critique of transpiled output is that it's ABL in Java clothing: procedural patterns your Java hires won't recognize or want to own. Marcus takes a different path: the source becomes a human-approved specification, and the rebuild is native, documented Mendix — judged app by app against its test evidence, not by a conversion percentage.

We're mid-contract with Progress. Doesn't that lock us in?

No — it's actually the right window. The old system keeps running, fully licensed, while apps migrate wave by wave and cut over only when you accept them. You exit at renewal from a position of strength instead of negotiating with no alternative. We'll flag anything in your license terms (especially ISV source rights) that affects sequencing during the assessment.

Our power users fly through the character screens. A web UI will slow them down.

For a few weeks, honestly, yes — muscle memory is real. We mitigate it deliberately: keyboard shortcuts, inline editing in data grids, focused task pages designed in a UX-parity workshop with your actual heads-down users. And the trade is bigger than one desk: everyone who is not a power user — new hires, managers, the warehouse, anyone on a tablet — stops paying the green-screen training tax.

We have forms and reports that must print exactly right.

Named, scoped, and priced per document — not hand-waved. Operational reports usually rebuild better as filterable pages and exports; the genuinely layout-critical documents (invoices, compliance forms) are produced through our in-process PDF generation and proofed against your current output as part of acceptance.

Who owns the result?

You do — full stop. Standard Mendix applications in your repository, with documentation, test suites, and the audit trail included. Cancel any time and your team or any Mendix partner continues in Studio Pro exactly where Marcus left off.

10 · Next step

Send us your .df and three screens

Export your data dictionary and pick three representative .w screens — a workhorse entry screen, a lookup, and your hairiest one. Marcus reads them and returns the scored specs — what carries forward, what changes, and a fixed price to rebuild each — inside two weeks. Judge the approach on your own application, not on a slide.

Marc Lehane · Solutions Architect / CTO, Kinetech · [email protected]

Kinetech Cloud · San Antonio, TX · How we deliver · Live Build Console

123456

You are reading Your platform — deeper detail on step 4, Why now?.

That is the pressure. What it costs to answer it is not one number — three things move it.

All platformsNext — What does it cost?