Current project has a ridiculously high amount of formal requirements

Two of them literally are
• the system shall have an “off” state
• the system shall be in the “off” state if no power is supplied

Boy do I have just the solution for you!

To be fair, an "off" state could also be that no clock is fed into the process or whatever without cutting off the power etc. Be happy that they have detailed specs instead of adding specs each week which were important to know from the beginning :(

@NEMESISprj we have several hundred requirements, it's insane

My boss said as well that we'll aim for like 40...

@LotsOfCaffeine oh damn, I didn't expect that. Big project or just big spec?

@NEMESISprj it's no small project but I don't know what they were thinking when writing all of that up

@LotsOfCaffeine You better have a traceability and coverage matrix.

@LotsOfCaffeine Oh, and the "off state without power" requirement isn't as stupid as it looks. It also means that you don't need to design in any battery backup power.

If you have requirements for the transparency time during power interrupts, that will only need sufficient capacitor buffering.

@Fast-Nop I mean the hardware interface were using is already defined, we're not handling power management ourselves

Hell were the software department

@LotsOfCaffeine I already had HW requirements like that with impact on software, e.g. how long the system will work with a power cut, and how to save data. The software had to switch off some external loads to have enough time for writing to flash.

@Fast-Nop yeah I get that, but considering that as of right now the system doesn't save any data to disk it

I just don't get why requirements like these were made, did the authors (I don't now who they are or in what department they are) aim for a high score in the amount of requirements that they hand over.

Like others (which I won't go into, because company internal) repeat themselves like 2-3 times

That switch is the wrong approach, you're going above and beyond the requirements. The spec doesn't say you need a way of switching the power off, just that your machine shouldn't work without it 😁

Upon reading that closely... the off state is in singular. That would mean the machine has only one off state, which in turn implies that its outputs must be in a well-defined state upon power-off.

Sounds trivial, but the devil can be in the details. Just think of a solenoid with two stable mechanical states instead of a spring mechanism, such as in latching relays.

Looks like they outsourced their last project to the cheapest contractor available and now are somewhat overcautious...

@Oktokolo oh nah, we don't outsource like that
I honestly don't know why they went overboard like that

Do you have compliancy requirements for certifications or regulatory bodies? This smells like that.

If so, these are *good* things. It leaves absolutely ZERO room for doubt, which is always a great property of software.

If this is just some project manager thinking he's hot, then yes I agree this is absurd lol.

Somebody's new to the whole "design" thing...

If the off state is functionally equivalent to powered down and it's achievable by input then you've already excluded most modern phones.

@homo-lorens how so? Powering down I can do with a long press and a tab on any phone I had. Unless you mean that there is a way to cut off the power than yeah just like laptops etc. They don't come with detachable batteries anymore.

@hjk101 Turning off doesn't always power down the device; there are wake-up timers, the cell beacon doesn't turn off until you remove or drain the battery and I think low powered sensors can also keep running.

@homo-lorens beacons and sensors no, timers vary. I still think it counts as in off state. My PC with WOL and other active boots components won't count either otherwise. I just see timers as another input/subsystem to put it in ON state again. Just like the button (needs long press to start so no simple switch). For all intents and purposes the system is off and does not perform tasks until it is on

Remove the small battery on the mainboard then and if they wonder about wrong date/time settings and funny error messages on startup then tell them "as required by the spec"

@hjk101 A state where the device can behave differently than it does without power (excluding the ON signal) isn't functionally equivalent to power-off state. I would readily disqualify WOL-capable devices.

@hjk101 Depends on whether you consider the timer/trigger a part of the system. A hardware expert probably wouldn't and then the description really does fit just about every device.

Nice, free unit test templates.

"the system shall be in off state when no power is supplied"

boy, dodged a bullet there