Comment by pixl97
2 days ago
Ya I don't think I'd let you in two miles of a system like this.
Replacing legacy stuff always expands in scope far beyond the initial changes.
When you have to come back and add wait() entries in your new program because it spits data back faster than the old program ever could which then causes peripheral devices/drivers to crash which then pulls a dev and testers off something else important for days figuring out what kind of fresh hell is occurring is just status quo for ancient systems.
idk much about dev much less legacy / enterprise dev but it seems like an A/B test type situation where you have 90% of the users running the legacy code and the remaining 10 on a new implementation would be feasible any idea why this is the case?
This is what happens. All that testing with the required stakeholders takes way way more time than you'd expect.
Gets even more fun in .gov where the work can change significantly at particular times of the year.
Had one piece of Windows software required by the State of Texas used at year end like once a year. Seemingly nobody realized windows updates had stopped it from working until a few weeks before the deadline. I had to setup a box without updates for it to run for my customer. Lead to a lot of panic around the state.