Welcome • User Guide
ToS • Privacy • Canary
Donate • Bugs • License

©2025 Poal.co

400

Just like with the vaccine, you can't expect to completely re-write a 60-million-line COBOL codebase in just a few months. Unless he has some ungodly strange tactic I'm not familiar with, this can't happen.

Just like with the vaccine, you can't expect to completely re-write a 60-million-line COBOL codebase in just a few months. Unless he has some ungodly strange tactic I'm not familiar with, this can't happen.
[–] • 6 pts 9d (edited 9d)

You don't need to rewrite the whole thing, or read any Cobol. Collect the requirements, write the code.

It shouldn't be too difficult, the actual rules on dispensation are relatively simple. The code to send checks to a printer or ACH processor is simple and ubiquitous. Don't let anyone fool you, just because it's government and important doesn't mean it's particularly special.

A company I worked for had been running on AS400 for decades, but we were able to move to a SQL based backend with a web app on top of it, and we were in UAT within 1 year. Believe me, the financial data we dealt with was far more complex and convoluted than Social Security, with exceptions to every business rule, exceptions to exceptions, and a minor fuckup could expose us to massive legal risk.

This shouldn't be difficult. If SSA can define the business rules and RBAC, this should take less than a year with the right technical resources and a good project team.

[–] • 1 pt 9d

^ this

This lame excuse for not rewriting code gets used too often by people who don't know how to program. COBOL is not ancient magic. You don't need to touch it if you simply build a new system from the requirements as was said above. The real reason they don't want to do that is that a rewrite will expose how little they actually know about the current system and the actual rules and logic that govern it. A new system will shine light on the fuckery that was put in the original system to do the things they want hidden. A good team of non-pajeet programmers (i.e. straight White men) can replace this old code in a year or two, but that would, again, expose the fuckery in the system. They will live with the COBOL because it hides the bodies.

[–] • 3 pts 9d

I disagree with some of the conclusions the writer comes to. Smart, actual software engineers can learn COBOL very quickly, in a week to 2 weeks. It is far less complicated than modern software languages. Yes, I’m sure a system written 60 years ago and cobbled together with other programs many times per decade is highly complex, however, a highly capable and very intelligent software engineer could figure it out. It would take more than a few months, obviously, and you won’t get anywhere with pajeet and DEI coders.

My last job in IT involved working with Medicare/Medicaid claims. We were sent claims and our system could only read the first 80 fields of a claim line. It took a few months to find someone at the state to root out an old manual and she read it to me over the phone. We then knew what the remaining 250+ fields were and how to read them. They’d forgotten how the data was used. Old COBOL system and from a single phone call I solved our problem. I then wrote up a guide and sent it to the state for their developers and other contractors to reference.

[–] • 3 pts 9d

Chainsaw Al lives again.

[–] • 3 pts 9d

I was thinking perhaps he believes Grok can do it. I don't think so. But maybe we'll find out.

[–] • 1 pt 9d

Anglin said it best - Musk is a fraud and a retard.