Re:Most rewrites serve the engineer, not the business
by wl on Wednesday July 01, 2026 @06:11PM
(#48753804)
The big exception to this I've seen: The functioning bit of software was written in an oddball language that was already niche decades ago and the pool of developers who are competent to work on such a codebase is exceedingly small.
[ Parent
| on HN ]
Re:Most rewrites serve the engineer, not the business
by darth_avocado on Wednesday July 01, 2026 @06:55PM
(#48754220)
I would classify them in the “ existing languages/frameworks getting obsolete” bucket.
Re:Most rewrites serve the engineer, not the business
by jamesfinlayson on Thursday July 02, 2026 @01:44AM
(#48756996)
Yep - I was at a company that replaced something super obscure - I suspect the number of people in Australia that knew the tech was like 10 to 20. The tech was so obscure I don't think there was any path out other than a full rewrite.
A company I worked at decided to rewrite a Fortran system - I suspect that if they wanted to avoid a rewrite they could have started breaking bits off into C and mixing the languages at the linking step... but I understand their desire for a clean break - a rewrite was going to take years but a migration would have taken even longer and been even harder due to the lack of Fortran developers.