The team working on this project has decided that since the coding for the method involving human verification was basically done, they would let it stand and make notes of my much simpler method for "future development and enhancement". Fine, whatever, if it works it works, and the answer to a lot of the questions about timeouts and escalations is "there's no excuse for the person whose job it is to respond not to respond, if they fail to respond that is poor work performance" and since that is not me, I am ultimately fine with whatever as long as the failure state ends up being "we think there's an event when there is not" (results: extra samples to be discarded, wasted time and materials) and not "we think there is not an event when there is" (results: regulatory violation).
But also the plant manager was brought in to the meeting today, and he basically confirmed my theory about the genesis of the project: because there was one (1) instance about two years ago of someone not getting a phone call, and that led to a contaminated sample, and that led to a regulatory violation, there must be a Computerized Process to make sure it is never ever ever missed again. The goal is apparently that, when the event is detected and validated, a big screen will light up indicating that The Event Is Happening and listing the samples to be collected.
I have seen some Solutions In Search Of Problems in my time, but this is a doozy.
We already have a paper checklist. We already have a flowchart. We have a giant monitor on the wall that shows the basic state of the process control system, which conveniently indicates whether the event is happening or not. Nothing going on here will prevent the mistake that was actually made - because of the missed phone call, at the end of the event, the sampler was not shut off on time. We are spending hundreds of man-hours on a computer project that will not solve "someone was working in a noisy area and wearing ear protection and didn't hear the phone ring" (or "I was at lunch and left the phone on my desk" or "I left the phone in the truck so I didn't accidentally drop it into the industrial process" or whatever actually happened.)
...and yes, I have expressed this to the plant manager, who countered with "I am responsible for all regulatory violations and the only acceptable number is zero, therefore I must Do Something About It and this is what I am choosing to do."
I hate to be the one to tell him but any requirement that a collection of humans never collectively make even a single error is doomed to failure...
But also the plant manager was brought in to the meeting today, and he basically confirmed my theory about the genesis of the project: because there was one (1) instance about two years ago of someone not getting a phone call, and that led to a contaminated sample, and that led to a regulatory violation, there must be a Computerized Process to make sure it is never ever ever missed again. The goal is apparently that, when the event is detected and validated, a big screen will light up indicating that The Event Is Happening and listing the samples to be collected.
I have seen some Solutions In Search Of Problems in my time, but this is a doozy.
We already have a paper checklist. We already have a flowchart. We have a giant monitor on the wall that shows the basic state of the process control system, which conveniently indicates whether the event is happening or not. Nothing going on here will prevent the mistake that was actually made - because of the missed phone call, at the end of the event, the sampler was not shut off on time. We are spending hundreds of man-hours on a computer project that will not solve "someone was working in a noisy area and wearing ear protection and didn't hear the phone ring" (or "I was at lunch and left the phone on my desk" or "I left the phone in the truck so I didn't accidentally drop it into the industrial process" or whatever actually happened.)
...and yes, I have expressed this to the plant manager, who countered with "I am responsible for all regulatory violations and the only acceptable number is zero, therefore I must Do Something About It and this is what I am choosing to do."
I hate to be the one to tell him but any requirement that a collection of humans never collectively make even a single error is doomed to failure...