Nine Products — Nine Markets — Nine Separate Licenses
Nine-division compiler portfolio that converts COBOL, RPG IV, PL/I, HLASM, CL, JCL, and BMS into modern C++, Java, and Bash. Not a wrapper. Not an interpreter. Not an emulator. A full source-to-source compiler built from first principles.
Deterministic. Non-AI. Auditable. Every line of output traces back to the source.
Nine Products — Nine Markets — Nine Separate Licenses
$3 trillion in daily transactions. 84,000 unfilled positions. A talent pipeline that no longer exists.
Average COBOL developer is over 60. Retiring at 10%/year. No next generation being trained. When they leave, institutional knowledge leaves permanently.
Manual modernization: $100M–$500M, 5–10 years, fails more than half the time. TSB Bank: $330M. SSA: $677M. Most never finish.
Government and regulated industries require source-level control, audit trails, and agency-specific compliance. Off-the-shelf tools don't provide this.
100,000+ businesses worldwide run RPG IV on IBM iSeries. No commercial compiler exists. They are completely stranded.
Level 1: NIST ANSI standard. Level 2: Full IBM v6.4 specification. Level 3: 98,662 real-world programs from 53 repositories.
Nine independent products. Nine separate licenses. Nine non-competing customer segments.
COBOL → C++17 / Java. Byte-perfect behavioral fidelity. CICS, DB2, IMS, JCL, COPY REPLACING. 220/220 tests.
Per-agency compliance modules. Source-only delivery for air-gapped environments. SOX, PCI-DSS, HIPAA, IRS 1075, CMMC, FedRAMP, DISA STIG.
RPG IV → Java. The only commercial compiler in this space. 303/303 tests. Fixed, free, and full-free format.
The only commercial solution for the assembler problem every COBOL migration hits. 250+ tests. 8 IBM catalog macros.
Binary → SQL migrator. EBCDIC, packed decimal, zoned decimal directly to PostgreSQL/MySQL. PII detection. SHA-256 integrity.
PL/I → C++. PL/I grammar is 3x more complex than COBOL. 156/156 tests. ON conditions, ENTRY points, structures.
The only commercial CL migration solution. 46/46 tests. Dual target: Bash + Java. Completes the full AS/400 picture with WebXRPG.
JCL → modern shell scripts. Preserves job step sequencing, DD file allocations, and execution logic.
Extracts 3270 screen layouts to JSON field maps. Pluggable UI hooks replace terminal I/O with any modern frontend — web, REST API, or JSON socket. Complete 3270-to-browser migration path.
No IEEE 754 floating-point in the financial path. $1,234,567,890,123.45 + $0.01 = exact to the penny. Every time.
18 significant digits. Every field stored as a scaled integer internally. No floating-point. No conversion error.
34 significant digits. IEEE 754-2008 hardware decimal. Trillion-scale balances, 30-year compound interest, actuarial tables. Optional compile-time flag.
MathContext.DECIMAL128 everywhere. Every operation uses BigDecimal. No toDouble() in any arithmetic path.
Native C++ with zero GC pauses. 2–8 MB per program vs 50–200 MB JVM. <5 ms startup vs 800 ms + JIT warmup.
WebXCOBOL: 1.42s — 7.04M rec/s
GnuCOBOL: 2.18s (1.5x slower)
Java 21: 4.61s (3.3x slower)
AWS Blu Age: 5.2–8.1s (4–6x slower)
WebXCOBOL: 3.87s
Java 21: 8.94s (2.3x slower)
IBM z17 DFP silicon: ~5 ns
WebXCOBOL + BID128: ~20 ns
Java BigDecimal: ~80–150 ns (4–7x slower)
C++ double: ~2 ns but wrong
AWS Blu Age: $38,111,768
WebXCOBOL: $100,000 one-time
Savings: $38M (99.7%)
Every acquisition below was purchased because customers cannot leave the mainframe. We give them the exit.
| Capability | WebXCOBOL | AWS Blu Age | Micro Focus | IBM COBOL |
|---|---|---|---|---|
| NIST 100% conformance | 426/426 | Unknown | Yes | Yes |
| Native binary output | C++17 | Java/JVM only | Partial | No |
| Runtime license fees | $0 | $/core/hr | Per-seat | Per-seat |
| GC-free operation | Yes | No | No | N/A |
| Startup latency | < 5 ms | 300–800 ms | Partial | N/A |
| Data lineage (SOX) | Built-in | No | No | No |
| RPG IV compiler | 303/303 | No | No | No |
| HLASM compiler | 250+ | No | No | No |
| Auditable source output | Yes | No | No | No |
Micro Focus → OpenText (2022): $6.0B • CA → Broadcom (2018): $18.9B • Kyndryl spinoff (2021): $19B
Three-layer audit architecture. SHA-256 manifest in every binary. CycloneDX SBOM per compilation. Zero external dependencies.
Financial field audit ledger with data lineage tracing. SHA-256 binary provenance. ISO-8601 timestamps.
Encryption hook architecture, PAN masking, auto PII field detection across 13 categories.
17 controls mapped (SI-7, SA-11, CM-14). ATO-ready artifact package. Air-gap deployable.
PHI field tagging, encryption hooks, audit controls. De-identification support for 13 PII categories.
SI-7 integrity controls. Agency ATO pathway documented with SSP and POA&M templates.
CUI flow control, build environment documentation, component inventory via SBOM. C3PAO pathway.
Availability, integrity, and confidentiality controls supported by audit manifest architecture.
ICT risk mapping, field-level lineage, CycloneDX SBOM. Zero transitive dependencies.
Data lineage graph for right-to-erasure impact analysis across transpiled code.
Auto-generated per compilation. Full NTIA minimum elements. CycloneDX 1.4 JSON format.
Extended artifact set. C3PAO assessment.
SHA-256 prerequisite met. NVLAP lab required.
3PAO documentation. Agency sponsor required.
Tool qualification per DO-330. FAA DER required.
FTI chain-of-custody audit trail. Agency engagement required.
Classified manifest field schema. Security clearances required.
Source-to-source transpilation with header-only runtime. No VM. No interpreter. No middleware.
Schedule a technical briefing. We'll transpile your code live and show you the output quality your auditors will approve.
Request a Briefing