It's a lot easier to test a conversion than something totally new. You can run the same reports on both and confirm you get the same answers. This likely means that the testing can be automated. Automating the coding part of the conversion will be a disaster. For one thing, it will not really be Java. It will be COBOL with a mask. Ever read something translated by someone who wasn't fluent in both languages (a menu, for example)? It will probably end up being less maintainable than what you started with. --Alan -------- Original Message -------- SUBJECT: Re: [Hidden-tech] Rewriting SSA code in months - possible? DATE: 2025-03-29 17:43 FROM: Tom Kopec via Hidden-discuss <hidden-discuss at lists.hidden-tech.net> TO: hidden-discuss at lists.hidden-tech.net On 3/29/2025 1:12 AM, AI Zuckerman via Hidden-discuss wrote: > Is it possible to handle this mammoth job in a month? In a word, no. both my wife AND my daughter have deep experience in implementing and converting large systems like this. Their first question: "Where's the test plan?" I get the feeling their plan is to crock up some AI to convert everything. That's going to result in impossible-to-debug bugs. go-live will be a total cluster. Just MHO. ...tom -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.hidden-tech.net/pipermail/hidden-discuss/attachments/20250329/0112e269/attachment.html>