Economists are now admitting that they were wrong about globalization

Re:Economists are now admitting that they were wrong about g... by generatorguy on Friday October 25, 2019 @05:54PM (#21359619)

Oracle / IBM had an existing working payroll system to copy. All they had to do was reimplement it in modern hardware and software. Some government employee in Miramichi might not have been happy because oracle didn’t listen to what they thought they wanted but if the contract has been to deliver a system that produced the exact same output as the old one oracle/ibm would have no excuse for not having a successful project.

It is astounding nobody was fired for this.

[ Parent | on HN ]

Re:Economists are now admitting that they were wrong about g... by jeffdavis on Friday October 25, 2019 @07:04PM (#21360093)

I'm not so sure it's easy to represent the exact behavior of a complex system.

What would your estimate be to: "rewrite facebook in language L, with the same behavior, bugs and all".

[ Parent | on HN ]

Re:Economists are now admitting that they were wrong about g... by pas on Friday October 25, 2019 @08:47PM (#21360734)

Doing an incremental approximation is very much possible. And it's up to the customer to say when to stop. And also to judge the sufficiency of the speed of convergence.

[ Parent | on HN ]

Re:Economists are now admitting that they were wrong about g... by nradov on Saturday October 26, 2019 @12:32AM (#21361734)

An incremental approximation is unacceptable for payroll. It legally has to be 100% accurate.

[ Parent | on HN ]

Re:Economists are now admitting that they were wrong about g... by pas on Sunday October 27, 2019 @10:42AM (#21369981)

No, you misunderstood. It means you have two systems, Old and New. You use Old and start development of New. And you do the payroll on both, and if there's a difference you debug that and modify New. After a while it's good enough. And you stop using Old.

[ Parent | on HN ]