4 ms·
You are failing to engage with the content here. Work itself is behaving like a complex system with bottlenecks and constraints and all that. The deadlines are
by crdrost 4y ago
You are failing to engage with the content here.
Work itself is behaving like a complex system with bottlenecks and constraints and all that. The deadlines are being shown to expand to the time allotted, in a similar fashion to how opening a new lane on a highway often does not improve traffic and might even make it worse.
In complex systems, the “greedy algorithm” (gradient descent) fails. For example, you might see some welders playing cards, and yell at them to go be doing work, (greedy algorithm: any inefficiency you see, correct it immediately) but this might cause huge delays for your customers rather than making things faster. How? Well it comes down to why were they playing cards. If it truly had to do with them being lazy, that's one thing. The Union rep might even back you up and yell at them for you. But they might also be implementing a low-latency subunit. They don't have work to do right now and it's important that the moment they have work to do, they can start it in an instant. You send them all to look productive God knows where, then when work comes in folks have to hunt them down and get them back here. Or, the card playing could implement a backpressure mechanism, “yeah we know the QC folks who will handle these are still pretty swamped with the backlog...”
So if you give me a linear circuit comprised of resistors and capacitors and inductors, and you propose increasing the voltage on the lines that feed it, I know that other voltages seen across the circuit will scale proportionately—this is your “20% more.” and same with “if I reduce this resistor's resistance, more current will flow over there” etc. But once you throw a diode or transistor in there, all bets are off! “No, when you tweak this the diode goes from being forward biased to being reverse biased and NO current flows over there.”
So it matters terribly much whether work is a simple linear system or a complex of feedback loops.