If IBM does not their new COBOL Devs to make the move to different fucking language then they are total idiot's.

Crack is one hell of a drug

Rule number 1 for large projects: you don't rewrite. Rule number 2: you fire everyone who doesn't get this.

@Fast-Nop and that's how your project dies?

@aggelalex Rewrite is how your project dies, and the whole company along with it. You invest a lot of money into something that has no benefit for the customers so that they won't pay for it.

The next mistake is the delusion of grandeur that the new product will be better. Much of the quirky code in the old one is from field testing and fixing corner case bugs, and your new product won't have it yet, so it will be actually worse.

On top of that, your dev team is distracted from the current product line so that this falls behind.

A rewrite is a great idea to tank a company. That's why you fire people who don't get this. One, they won't harm your company. Two, with any luck and if hired, they will harm your competitors with their ideas.

@Fast-Nop well then _good luck trying to find COBOL devs_. If you think paying enormous salaries for Devs that code in ancient tech will still be more profitable to the company than rewriting/replacing, then you need to think more about the long run.

@aggelalex There won't even be a "long run" because the company goes out of business. Rewrites are a major mistake because they tank the whole company.

Unless of yourse you find rewriters who work for free for several years and only demand their pay afterwards once the new system is at least as good as the old one.

@Fast-Nop we are not talking about some small company or whatever. We are talking about a government. A government's part cannot go out of business as easily. But it can be in loss of workforce. Right now they are looking for COBOL devs. The next ten years what are they gonna do? Still be looking for COBOL Devs? Or are they going to pay people millions and billions each to continue development?

@aggelalex It's not difficult to find Cobol programmers because people can LEARN it. Even if the gov has to pay like 50% above market rate for the maintenance devs, that's still peanuts compared to a rewrite.

What the gov might do however is changing the spec by drastically simplifying the laws that are the basic spec input. That could work because the new system could be much less complex than the old one. But good luck trying that in a democracy, given that democracy is driven by compromise.

Nah, what I see here is the basic "rewrite from scratch" that most programmers always favour, and which is a big reason why programmers mostly don't run companies. Because they'd run them into the ground anyway.