Uniface. Magic. Clarion. Gupta. Informix 4GL. Visual FoxPro. The vendor was acquired, went quiet, or is gone; the last developer who truly understands the app is leaving — and the app still runs the business. Marcus, Kinetech's AI Mendix developer, reads these platforms' own source formats and rebuilds each application as modern Mendix: gated by your approvals, proven with published evidence, delivered to your own repository.
01 · The pattern
Every 4GL on this page made the same promise in the 1990s: build business apps fast, on our runtime. The apps delivered. The runtimes became one-vendor islands — and an application's lifespan should never equal its vendor's attention span.
| Platform | Steward & state | The pressure | What Marcus reads |
|---|---|---|---|
| Uniface | Rocket Software — maintained, niche subscription | Small ecosystem, roadmap anxiety, heavy recurring fees for a stagnant runtime | Component/model text exports |
| Magic xpa / uniPaaS | Magic Software — xpa maintained | Older uniPaaS/V8–V9 estates have no practical upgrade path — frozen where they stand | XML repository export |
| Clarion | SoftVelocity — releases continue from a very small vendor | A one-vendor ecosystem with a shrinking community carrying mission-critical vertical apps | TXA/APP template exports + DCT dictionary |
| Gupta / Team Developer | OpenText lineage | TD 7.4 sustaining support ends July 2026, per the published lifecycle; users openly ask about discontinuation | .apt/.app text outlines |
| Informix 4GL | IBM — Continuing Support for all Informix versions withdrawn April 30, 2025, per IBM's support pages | Frozen language on a database exiting standard support — deadline-driven | .4gl + .per plain text |
| Visual FoxPro | Microsoft — support ended 2015 | Runs only via 32-bit compatibility; every Windows release is a gamble | .prg text; .scx/.frx are readable tables |
| 4D (binary-mode) | 4D SAS — vendor healthy | Only legacy binary-mode solutions belong here; modern project-mode apps are fine where they are — we'll say so | Project-mode text; binary-mode via structure export |
02 · The differentiator
The common thread across every platform above: the application's design lives in text or readable structured formats — data dictionaries, template exports, repository dumps, plain-text procedures. Marcus reads those artifacts and derives each app's full functional specification, without workshops asking users to remember how a 1998 screen works.
| What the 4GL stored | What Marcus extracts | Where it lands in Mendix |
|---|---|---|
| Data dictionary / schema | Tables, fields, types, keys, relationships, validation rules, labels | Domain model: entities, associations, validation |
| Screen / form definitions | Layouts, fields, lookups, grids, keyboard flow | Pages with data grids, reference selectors, task-focused layouts |
| Procedural 4GL code | Business rules, calculations, batch routines, report logic | Microflows and scheduled events, documented |
| Menus & navigation | Screen flow, role-gated entry points | Navigation model and page access |
| Users, groups & gating | The security model as actually implemented | User roles and entity access rules, verified by automated per-role tests |
| The data itself | Profiled as it actually exists — including the quirks | Migrated to standard SQL with exceptions surfaced for your decision |
Why this matters: nobody staffs a practice for a platform with a four-figure customer count — which is exactly why these estates are stranded. An AI developer doesn't need a practice. If the artifact is readable, Marcus reads it, and the rebuild runs through the same evidence-gated process as every other platform we migrate.
03 · How a migration runs
The four approval gates that govern every Marcus engagement are held by you:
The extracted specification — what the app does, who uses it, what carries forward — approved by you before a line of the model is written.
The translated security model in plain English — often the first time it has ever been written down — for your sign-off.
Modern, branded design directions presented before pages are generated. You pick what your users land on.
The app reaches you only after all quality gates pass; it cuts over only when you accept it, evidence in hand.
04 · The date
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.
| One application | Complete artifact pack → UAT-ready |
|---|---|
| Compact, readable source — Uniface, Magic, Gupta/SQLWindows, Informix-4GL, Clarion | 2–4 weeks |
| Departmental, readable source — 2–5 integrations, real printed output | 4–6 weeks · ~30 days typical |
| Binary-only or partially binary — compiled-only estates, lost source | add 2–3 weeks for reverse-engineering |
| Load-bearing — several applications on one database, no surviving developer | 8–14 weeks |
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.
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.
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.
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.
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.
These are the rescues, and honesty about the variance matters more here than a headline. Where source or a repository export survives, this runs like any other migration on this list. Where it does not, the assessment's first job is to establish whether the application can be specified at all — from the database, from the running screens, and from the people who use it. That answer arrives inside the assessment, before you commit to a rebuild date.
05 · Why Mendix
The mistake wasn't choosing a 4GL in 1997. It was ending up somewhere an application's future depends on one company's quarterly priorities. The rebuild should fix the trap, not re-enter it.
06 · Straight talk
| Area | The honest position |
|---|---|
| Platform quirks | Each 4GL hides its own surprises — index corruption and century-ambiguous dates in old FoxPro data, repository idiosyncrasies in Magic and Uniface, template-generated code in Clarion. The assessment exists to find them before they're mid-project discoveries; the ones we find are named in the spec you approve. |
| Not every app should migrate | Some of these applications should be replaced by off-the-shelf SaaS in 2026 — a vertical package or a $50/month tool. Where that's true, the assessment says so and we lose the line item. You rebuild what differentiates you, nothing else. |
| Keyboard-driven UX | Power users on keyboard-first screens will be slower for a few weeks — we say that plainly. We run a UX-parity workshop per app: shortcuts, inline grid editing, focused task pages. Everyone else — new hires, occasional users, anyone on a browser or phone — gets faster on day one. |
| Print fidelity | Operational reports rebuild as pages and exports; layout-critical documents (invoices, forms) are produced through our in-process PDF generation and proofed against your current output as part of acceptance — scoped per document, never hand-waved. |
| Data migration | Marcus profiles the data as it actually exists and surfaces exceptions as decisions — mapped, defaulted, or archived. The original data files are archived read-only as the system of record; nothing is destroyed. |
| Deal size honesty | These apps range from a single tracking tool to a company-running vertical system. The assessment prices each one from its actual extracted complexity — small apps get small quotes. |
07 · Quality & security
Watch real build evidence land on the live Build Console, and see the delivery lifecycle at our interactive explainer.
08 · Engagement & pricing
Design export and automated read of your application(s): a scored inventory (rebuild / retire / replace-with-SaaS), complexity and risk flags (data quality, source rights, print-critical documents), and fixed-bid quotes for the rebuild.
Credited toward onboarding if you proceed.
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.
Small apps get small quotes; the vertical system gets a staged plan.
For a portfolio of legacy apps: 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
Some of these vendors do keep shipping — we say so in the table above rather than pretending otherwise. The question isn't whether the vendor releases a maintenance build next year; it's whether the ecosystem around your app — developers you can hire, integrations, OS compatibility, a second opinion when something breaks — is growing or shrinking. For every platform on this page, that answer points one direction. A roadmap from a vendor is not a labor market.
Bus-factor math. If one person's departure would leave your company running software nobody can safely change, that's a standing business risk — the kind auditors and insurers increasingly ask about. Migrating deliberately, while the app works and its expert is available to validate behavior, is the cheap version. The expensive version starts with an OS upgrade that breaks something, or a resignation letter.
Sometimes, at scarcity pricing — and then you've re-created the exact single-point-of-failure you were worried about, one retirement later. The pool only shrinks. A Mendix rebuild converts the app to a mainstream, hireable skill set and ships with documentation and tests, so no future hire ever holds the only copy of the knowledge again.
That's the honest alternative, and for some apps it's right. Weigh what it means: hand-rebuilding the screens, validation, security, and reporting your 4GL generated for you, at day rates, with quality dependent on the individual contractors. Mendix regenerates that layer on a governed platform, Marcus does the reading and rebuilding from your actual source, and every release arrives with per-role test evidence. Get both quotes — ours is fixed-bid from extracted complexity, so comparing is easy.
It would sink a naive one — which is why profiling comes first. Marcus reads the data as it actually exists: orphaned rows, ambiguous dates, fields repurposed in 2004. The exceptions come to you as decisions in the spec — mapped, defaulted, or archived — before anything migrates, and the original files are archived read-only as the system of record.
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
Whatever your platform calls it — the data dictionary, the repository export, the TXA, the .df — send one design export from your application. Marcus reads it and returns the scored spec: what carries forward, what changes, what it costs, fixed. Inside two weeks, judged 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
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.