Modernization track · Intro · self-paced ~14h with labs (2 days as a cohort) · prerequisite: CS100 (or equivalent exposure to a mainframe estate); no programming
Decide how an estate should move — and prove the decision from the estate itself. Nine modernization styles in three families (rehost and COBOL-on-the-JVM; transliteration, tool refactor, AI-assisted and rewrite; package, hybrid and data-first), the factors that pick one per application, the lock-in each hides, the verification each needs, the cost shapes and the failure cases (TSB, Co-op/Finacle, Queensland Health against CBA). 28 short videos; every decision is made while looking at the real CardDemo estate on SteelFrame — its inventory, call graph, batch DAG, scheduler definitions and run data — never at a diagram alone.
Self-paced available Cohort
Open for enrollment now. The lesson videos for this course are in production — release order across the track: CS510 → CS520 → CS530 → CS521 → CS535 → CS500 → CS550 → CS525 → CS540 → CS545 — and every session below is published as its lessons and lab now and marked Coming soon until its videos are released.
Plans: Free individual at no cost, or Paid individual at $349 per course, one-time — code review, extra exercises, grading and priority support.
Every session ends in a lab on your own environment. Labs are machine-checked against expected outputs — you and your L&D team always know exactly where you stand.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Four videos. The drivers (cost, skills, agility, regulation) and the counter-evidence — MIPS growth, Gartner’s 70% of exits failing on overestimated GenAI, Kyndryl’s "hybrid dominant"; the nine styles in three families shown as three renderings of one CardDemo paragraph (CBTRN02C’s posting logic in BROWSE beside transliterated, framework-shaped and idiomatic Java); who sells what — AWS Transform, IBM WCA4Z/Bob, Google, Rocket, TmaxSoft, LzLabs, Fujitsu PROGRESSION, the SIs — named with the CardDemo artefact each would consume on screen, never endorsed. Live on the system: the COSGN00 sign-on and the nightly chain in SDSF as "the thing that moves".
LAB Classify ten real postings and ten vendor claims by style, family and the definition of done each implies, using the inventory_scan of CardDemo as the material.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Five videos on evidence. inventory_scan on CardDemo (44 programs with CICS/SQL/DL-I flags, 47 jobs, 65 copybooks, 21 mapsets); analysis_graph and the shared copybook as a seam (CVACT01Y read by batch and online); the batch DAG read from the JCL hand-off datasets, the GDG generations and the Control-M/CA-7 definitions in sched_list; analysis_metrics for complexity, dead paragraphs and dead members, and the build-gap list the scan cannot resolve.
LAB Produce the inventory pack for CardDemo — inventory.json, graph.json, metrics.json, dag.json, gaps.md — and reconcile its counts and edges against the harness’s own run.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Six videos. Decision factors (change rate, coupling, data gravity, regulation, skills, deadline — Fujitsu FY2030/FY2035 as the worked example) with the chain’s elapsed and RCs from JESMSGLG as factor data; rehost and COBOL-on-the-JVM in depth — what breaks (EBCDIC and packed data at rest, DFSORT semantics, DB2), what it costs per core, what it does not change, the IBM v LzLabs line; transliteration vs tool refactor vs rewrite with the lock-in each hides; AI-assisted conversion’s four tool families and the verification each ships, with explain_program checked against the real source; package replacement and hybrid coexistence with the SEA/PH/SG core landscape and the API seam shown live through cics_link.
LAB Build the style-decision matrix for CardDemo’s online and batch halves: every unit a style, every style a cited rationale, a lock-in note and a verification method — graded against the factor table.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Four videos. Slicing waves along the DAG with coexistence bridges where an edge must be crossed; data — copybook to DDL, the charset boundary, the control-total chain from trailer to report as the reconciliation spine, cutover and fallback; the verification plan per style (zero-tolerance bytes, record/replay, rules-file equivalence, behaviour contracts, continuous reconciliation) sized at 40–50% of effort, starting from a run_capture golden envelope.
LAB A wave plan and a verification plan for CardDemo’s nightly cycle, validated: every program in exactly one wave, no wave crossing an unresolved dependency, every wave with its goldens and verdict type.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Four videos. Cost shapes — per-core runtime (Rocket, Blu Age) vs per-line conversion vs rewrite effort, with the chain’s CPU and elapsed from JESMSGLG as the baseline; the lock-in map (emulation runtime, generator, verification product, cloud, assistant) made visible as a plain COBOL CALL beside the runtime call it becomes; the cases — TSB/Proteo, Co-operative Bank/Finacle, Queensland Health against Commonwealth Bank’s product-by-product success — and what the evidence says each lacked, with a CardDemo transaction driven under load on camera.
LAB A one-page risk register for your chosen style on CardDemo: ten risks with likelihood, impact, control and owner, built from the inventory’s gaps, the lock-in map and the cases.
Coming soon The videos for this session are being produced; its lessons and lab are published here now.
Five videos. IBM i estates (Silverlake-class RPG cores) with the same method on different artefacts — DSPFFD, CL, job queues on SteelFrame X, opening on the CardDemo copybook each mirrors; regional frames — Japan’s rehost/rewrite/rebuild under the 2035 deadline, SEA’s package-plus-hybrid, LatAm’s modernize-in-place; where AI helps (explain, inventory, tests) and where it does not (judging equivalence); writing the decision memo executives read, with its screenshots taken live.
LAB Capstone: the assessment pack and style matrix for a hidden CardDemo variant under your own user — inventory counts, DAG edges and wave membership checked against the harness’s analysis, the matrix against the factor table — plus the 30-question pattern-selection check and the executive memo.
CobolStack Certificate — Modernization Assessment (CS500) is issued from the lab harness’s own record of your runs — not from attendance. To earn it you must clear every gate below:
The certificate attests that you can assess a real estate from its own data, choose a modernization style per application with the lock-in and verification spelled out, and present the decision in a form a steering committee can act on — judged against the harness’s analysis, not an attendance sheet.
Your seat is a private, full z/OS-compatible environment with the AWS CardDemo estate and the analysis tools (inventory_scan, analysis_graph, analysis_metrics, scheduler definitions); SteelFrame X for the IBM i session in the browser — nothing to install, snapshot and reset per exercise. By the end of the course the lab harness will have verified: an assessment pack for a hidden CardDemo variant — inventory, dependency graph, batch DAG, complexity and dead-code findings, a wave plan validated against the DAG, a style decision per application with cited factors, lock-in and verification method, a risk register and a one-page executive memo — checked against the harness’s own analysis of the same estate.