Why we started writing this down

The group chat started as a place to complain about vendors. It became, without anyone deciding, the place where we asked each other the questions we were embarrassed to ask at work.

Why is my join duplicating rows. Is this result significant. What should I put on the y-axis. How do I know the pipeline actually ran. Should I use Spark for this.

The answers were the same every time, and every time they scrolled away.

Meanwhile the market for those answers had gone somewhere strange. Someone sent round a link to a course charging £1,400 whose syllabus was, verbatim, the section headings of the Postgres documentation. Someone else found a "bootcamp" that was a Notion page and a Discord. Some of it is good work by people who deserve to be paid. A lot of it is a reading list with a checkout button, and the pricing filters hardest on the people who can least afford a bad bet.

So: we are writing the answers down. Once, properly, in public, for nothing.

The plan is a crash course covering the whole shape of the job, a library of datasets with the mess deliberately left in, and some browser tools so you can poke at the data without setting anything up. All free, all licensed so you can take it and teach from it yourself.

We have no idea how long this will take. The first estimate was six weeks, which is how we know it will be eighteen months.

If it turns out to be useful, tell someone. That is the entire distribution strategy.