Resources
Research and references for modernization architects.
The catalog of legacy constructs that hide business rules, migration playbooks by stack, sourced benchmarks, a working DMN guide and the engineering notes behind our parsers. Free to read and cite.
Legacy anti-patterns catalog
25 legacy anti-patterns, mapped to modern designs.
From COBOL ALTER and GO TO DEPENDING ON to EJB entity beans and PL/SQL cursors: why each is dangerous during migration, how it is detected, and the before and after code.
Explore the catalogResearch and guides
Read the method, not the marketing.
Everything here is written by the people who build the parsers and run customer readouts. External figures are sourced; PYD ENTERPRISES figures are labelled.
DATASET · CC BY 4.0
Legacy Modernization Index
Footprint, cost of debt, public sector spend and migration consequences, from Vanson Bourne, CISQ, US GAO and UK regulators. No untraceable failure rates.
800B+
lines of COBOL estimated in daily production use worldwide
$1.52T
accumulated software technical debt in the United States
~75%
of the US federal FY2015 IT budget went to operating and maintaining existing systems
ENTERPRISE GUIDE
DMN 1.5 for architects
Decision tables, all seven hit policies with an interactive evaluator, FEEL basics, mapping COBOL to rows and validating in Camunda and Drools.
20 PLAYBOOKS
Migration guides
CALCULATOR
Discovery cost and timeline
Lines of code, languages and team size in; manual archeology cost and PYD ENTERPRISES timeline out.
56 TERMS
Glossary: COBOL to DMN
- REDEFINES
- A COBOL DATA DIVISION clause that gives a second description to the same storage as a preceding item at the same level. It is used for alternate record layouts selected by a type code and for reinterpreting bytes, so two names can alias the same memory without any tag recording which view is valid.
- COMP-3 (packed decimal)
- A COBOL numeric usage that stores two decimal digits per byte with the sign in the final nibble (C positive, D negative, F unsigned). A field takes (digits + 1) / 2 bytes rounded up, so PIC S9(7)V99 COMP-3 occupies 5 bytes; the implied decimal point V takes no storage.
- COMMAREA
- The communication area passed between CICS programs on LINK and XCTL, or carried across tasks by RETURN TRANSID. The receiving program addresses it as DFHCOMMAREA in the LINKAGE SECTION, finds its length in EIBCALEN, and is limited to roughly 32 KB; larger payloads use channels and containers.
- Hit policy
- The indicator on a DMN decision table that defines the result when more than one rule matches. Single-hit policies are Unique, Any, Priority and First (U, A, P, F); multiple-hit policies are Collect, Rule order and Output order (C, R, O), and Collect may aggregate with sum, min, max or count.
TRADE-OFFS
Compare approaches
Manual audits, LLM migration, transpilers, rewrites and rehosting, with honest notes on when each is the right call.
From the blog
Decompiler internals and field notes.
- Mainframe20 Oct 2026 · Nandini SubramanianBuilding control-flow graphs for PERFORM THRU, GO TO and ALTERHow PYD ENTERPRISES models COBOL paragraph ranges, fall-through, GO TO DEPENDING ON and ALTER in a control-flow graph precise enough for rule extraction.
- Modernization strategy1 Sept 2026 · Sahil SaxenaWhat the 2020 unemployment insurance crisis teaches about legacy rulesCOBOL took the blame when state benefit systems buckled in April 2020. The real lesson is about how quickly anyone could find and change the rules.
- Engineering14 Jul 2026 · Pranav KulkarniPL/SQL lineage: finding the side effects hidden in triggers and packagesStored procedures rarely do only what their names say. How PYD ENTERPRISES traces column-level lineage through triggers, package state and dynamic SQL.
- Modernization strategy9 Jun 2026 · Rohit MehraFixed-price discovery and the systems integrator margin problemWhy fixed-price modernization bids lose margin in discovery, worked arithmetic for a 2 million line program, and how a rule inventory changes the bid.
Newsletter · monthly
The Modernization Architect
One deep dive a month on mainframe decompilation, DMN extraction and legacy migrations, written by the engineers who build the parsers. No product announcements dressed up as research.
Stop spending six months reading legacy code by hand.
Run PYD ENTERPRISES on 100,000 lines of COBOL or Java in your private VPC and review OpenAPI specs and DMN tables tomorrow.