machines
The dragline
A metropolis for hydrocarbons
An oil refinery suggests the image of a metropolis for hydrocarbons, the pipe manifolds like expressways, the distillation towers like skyscrapers.
A haunting, syncopated music
"Rappers" on the roof of the electrostatic precipitator knock the accumulated dust free, letting it fall into the storage hopper. Each rapper is the size and shape of a baseball bat. Inside is an electromagnet that pulls a steel plunger upward, then allows it to fall again, producing a sharp knock. The rappers are energized at seemingly random intervals, producing a haunting, syncopated music. (The rhythm seemed more modern jazz than rap.)
One-machine policy
Today population forecasts are based on extensive and reliable data. However, no such demographic base exists for the world's growing population of machines and devices. Now may be the time to take machine demography seriously and enter into real discussions about machine population control.
An enormous machine
The couple of years in question here saw one of the largest bureaucracies anywhere undergo a convulsion in which it tried to reconceive itself as a non- or even anti-bureaucracy, which at first might sound like nothing more than an amusing bit of bureaucratic folly. In fact, it was frightening; it was a little like watching an enormous machine come to consciousness and start trying to think and feel like a real human.
Software that nobody wants
Finding value is the result of enabling individual and group-level discovery attempts. It's not the result of everyone following one leader's gut.
What just happened is a new software product/feature was created that no customer wanted. This happens way too often. In fact, most hyper important software projects that must be done by date certain or else, have deep flaws that cause some variation of this phenomenon, flaws that include:
- Not wanted - Company specified a solution to a problem that customers don't actually have
- No Rarity - Company is pursuing an iKnockoff of existing products. The market already has two scaled competitors with working solutions, customers naturally spend budget on products that are already successful to avoid risk
- Incorrect Packaging - Customers need a website, but the company created an iOS app instead
- Incorrect Pricing - Customers need SaaS pricing, but the company created a shrink wrapped, on-premise solution with CapEx and maintenance agreements instead
The 'date scrum' anti-pattern
Date Scrum is an R&D pattern where developers are asked to estimate software project requirements upfront for the entirety of the project. After the project is green lighted and the budget is set based on the final estimates, the team then holds daily scrums to status and manage risk as they “iterate” the solution toward the release date. To some, this approach is described as doing Waterfall in sprints.
The fundamental problem with Date Scrum is that the team is de-focused from discovering the best solution. Instead they are heavily focused on delivering Something™ by the Date™. Engineers are problem solvers, and if the primary problem becomes delivering Something™ that will pass QA by the Date™, they will, with enough pressure, solve that exact problem.