Industrial project, new machines to do new things, and the companies internal + one external team were tasked to buid an "enterprise grade" software to manage and control the process, show analytics, trends, etc., for everyone from the line worker to management, marketing etc.
They were working on it for a about a year and were maybe half-done, when the machines had to start the work, so they hired us (a third outside company) to "hack something up". It was a complex machine, configured internally, so basically it was just check if there was enough material (on a scale) on the "input" side, check that the output (a scale again) was not "full" (both configurable in a config file), trigger the machine to do its job, and save the diagnostic number (6 values) together with numbers for the scales to a database + grafana for visualizations. The machine was even smart enough to not start the work without material on the input and stop if the output was clogged, so even if we screwed it up, it wouldn't matter.
Two days later, we had a working prototype, a year later the internal team was still working on the software, and the "prototype" was still in production, then that project got dumped, and the ~100 lines of python on a raspberrypi (it was just a "prototype", just for a few weeks... yeah) are still running the machine... and still is.
The company bosses wanted it... the same ones that hired us for the "temporary solution"... we could charge a lot, because they were in a hurry, and couldn't find anyone else, so it was a good situation for us, but we were definitely cheaper than those two "enterprise grade" teams :)
Industrial project, new machines to do new things, and the companies internal + one external team were tasked to buid an "enterprise grade" software to manage and control the process, show analytics, trends, etc., for everyone from the line worker to management, marketing etc.
They were working on it for a about a year and were maybe half-done, when the machines had to start the work, so they hired us (a third outside company) to "hack something up". It was a complex machine, configured internally, so basically it was just check if there was enough material (on a scale) on the "input" side, check that the output (a scale again) was not "full" (both configurable in a config file), trigger the machine to do its job, and save the diagnostic number (6 values) together with numbers for the scales to a database + grafana for visualizations. The machine was even smart enough to not start the work without material on the input and stop if the output was clogged, so even if we screwed it up, it wouldn't matter.
Two days later, we had a working prototype, a year later the internal team was still working on the software, and the "prototype" was still in production, then that project got dumped, and the ~100 lines of python on a raspberrypi (it was just a "prototype", just for a few weeks... yeah) are still running the machine... and still is.