Thinking from the End State (2020)

Also known as backcasting:

Backcasting is a planning method that starts with defining a desirable future and then works backwards to identify policies and programs that will connect that specified future to the present.

https://en.wikipedia.org/wiki/Backcasting

I've found it a useful technique when trying to clarify vague or incomplete requirements.

Most people would call this having "vision". It can be pretty powerful, but it's often overrated.

A lot of people know what they want, or at least they think they do, if they gloss over all the messy details. What's really needed is the ability to break the vision into small, achievable steps that incrementally improve things along the way.

Stephen Covey coined this popularly in his "The 7 Habits of Highly Effective People".

Habit 2: Begin with the end in mind.

https://en.wikipedia.org/wiki/The_7_Habits_of_Highly_Effecti...

I think some of these examples are pretty easy to poke holes in, but the idea is pretty close to how I evaluate or plan around problems. Even when I’m making narrow and even urgent changes to something, I’m making a mental note if not a historical artifact tracking both my sense of the thematic problem and the general sense of where a holistic solution should go. I’m getting better at not getting too attached to the importance of every single long term goal—sometimes they’re a distraction, sometimes they’re wrong, often they’re both—but pretty much my entire professional methodology is a combination of trying to keep enough context to even think about end states, and using them to help guide short term decisions.

This was how we estimated project plans, at my previous employer (Japanese).

We always started from the ship date/spec, then backed through the various steps, until we got to where we were starting from.

It actually worked fairly well, for a strict, process-intensive waterfall angle, but I was seldom happy with the end result.

It sucks, when you get exactly what you asked for.

I’ve found considerable success, in “making it up, as I go along.”

Takes serious discipline, along with a willingness to toss out a whole bunch of work, and make sharp pivots.

I write about my process, here: https://littlegreenviper.com/miscellany/evolutionary-design-...

X is some state. What needs to be true or false for X to be true?

Someone might propose that Y is true. What minimum conditions exist for Y to be false?

And so on. Time does not need to be a variable in inductive reasoning.

I a big fan of the idea of future backwards thinking — living in the future, and looking back and seeing what we need to do to get there, and seeing what the blockers are.

This works very well in domains where there are lots of first order effects and where there are small measurable goals.

In ill-defined domains dominates by second or higher order effects and vicious feedback loops, I wonder if it works that well. Let’s say you wanted to live in a world where there’s no homelessness. How would you get there?

I liked this

> wouldn’t it be great if subway trains arrived every 2 minutes?

I encourage everyone with a dream like this to visit London where the Tube does come every two minutes, and at peak comes every 90 seconds (just long enough for a train to enter a station, open then close the doors, and depart). Just don’t hold the doors open on the train, there’ll be another one momentarily.

> How should a company reimbursement process work? Should the company require employees to write lengthy expense reports and get reimbursed, or just let them put stuff on a company card? Does it matter how many employees they have?

I believe the IRS wants receipts in the event of an audit, so reimbursement programs are set up to incentivize good documentation.

Or put simply: envision something and work backwards, step by step.

isn't that kind of.. debugging? i.e. digging up the reason why something had happened

although that one is about (past) reality, while OP one is about (speculative) future

maybe more like estimating of impact, requirements-wise, following a traceability chain.. what would break/change if this is changed?

It's a technique that I use often - "Start from the end"