COBOL Expert for Technical Review
Budget: $250 – $750 USD
Paid technical review — IBM Enterprise COBOL (z/OS) parity & dialect questions
I'm a founder building a tool that reads legacy COBOL, produces a plain-English specification of what a program does, and then checks that specification by independently re-implementing the logic and comparing the outputs against the original. I'm looking to pay an experienced mainframe COBOL practitioner for a one-off written technical review (roughly half a day, fixed fee).
The work: I have a structured list of 13 specific questions about where my approach could go wrong on real mainframe systems. They cover things like where GnuCOBOL silently diverges from IBM Enterprise COBOL on z/OS (packed-decimal precision, EBCDIC sort order, ROUNDED default mode, NUMPROC sign handling), how a client's exact compile options would be discovered in practice, how shops actually retain batch input/output files, and how much real batch logic is stateful or date-driven versus pure per-record calculation. I'll send the questions in writing. I'd like your written answers, plus — honestly — where you think the approach breaks down.
What I need from you (please address these in your application so I know it's a genuine fit):
Hands-on production experience with IBM Enterprise COBOL on z/OS — you personally wrote, debugged, or maintained it (not only GnuCOBOL, not only managing teams who did).
Real batch experience — JCL job streams, file-based processing, abend debugging.
Willing to tell me where I'm wrong — I want scepticism, not reassurance.
To apply, please briefly tell me: one legacy system you personally worked on (what it did, roughly when), and your fixed fee for a half-day written review. Generic applications that don't reference real z/OS experience won't be considered.
This is a one-off to start; if it's useful for both of us, there may be ongoing paid work.
I'm a founder building a tool that reads legacy COBOL, produces a plain-English specification of what a program does, and then checks that specification by independently re-implementing the logic and comparing the outputs against the original. I'm looking to pay an experienced mainframe COBOL practitioner for a one-off written technical review (roughly half a day, fixed fee).
The work: I have a structured list of 13 specific questions about where my approach could go wrong on real mainframe systems. They cover things like where GnuCOBOL silently diverges from IBM Enterprise COBOL on z/OS (packed-decimal precision, EBCDIC sort order, ROUNDED default mode, NUMPROC sign handling), how a client's exact compile options would be discovered in practice, how shops actually retain batch input/output files, and how much real batch logic is stateful or date-driven versus pure per-record calculation. I'll send the questions in writing. I'd like your written answers, plus — honestly — where you think the approach breaks down.
What I need from you (please address these in your application so I know it's a genuine fit):
Hands-on production experience with IBM Enterprise COBOL on z/OS — you personally wrote, debugged, or maintained it (not only GnuCOBOL, not only managing teams who did).
Real batch experience — JCL job streams, file-based processing, abend debugging.
Willing to tell me where I'm wrong — I want scepticism, not reassurance.
To apply, please briefly tell me: one legacy system you personally worked on (what it did, roughly when), and your fixed fee for a half-day written review. Generic applications that don't reference real z/OS experience won't be considered.
This is a one-off to start; if it's useful for both of us, there may be ongoing paid work.