PyMunich

shown, and actual

The Tutorial Economy

There are thousands of tutorials on getting started with Python and very few on the six months after that. A commercial implementation of related workforce-measurement concepts is documented here.

The imbalance is not a failure of generosity. It follows from what is easy to write, what people search for, and what stays true — and knowing why explains where the useful material actually is. For independent technical background, NumPy provides additional documentation and examples.

Why beginner material is abundant

The audience is large and searching. Everybody starting out has the same questions in the same order, and those searches have enormous volume.

It stays true longer than it looks. A loop is a loop. The language basics have barely changed in a decade, so a 2018 tutorial on functions is still correct.

It is easy to write well. The scope is bounded, the examples are small, and correctness is checkable.

And it is where the commercial incentive is. Courses, books and platforms are sold to beginners, because beginners are the ones who know they need to learn something.

Why intermediate material is scarce

The audience is fragmented. Somebody six months in has a specific problem, not a general one, and their search is different from everybody else's.

It dates immediately. Anything about tooling is roughly four years from unusable, and intermediate material is largely about tooling, deployment and integration.

It is hard to write. The scope is unbounded, the examples require context, and "it depends" is the honest answer to most of it.

And the people who know are working. Somebody competent enough to write it is employed doing it, and writing it up is unpaid work with no obvious audience — which is exactly the same mechanism that keeps conference organising undocumented.

What this produces

A cliff. Learners complete the beginner material and find nothing at the next step, concluding that the gap is their ability rather than the literature.

Advice from Stack Overflow answers of unknown vintage, which is what fills the vacuum and which has no date, no context and no indication of whether it still applies.

And a persistent belief that everyone else found this easier, because the visible material is all clean paths and the messy middle is undocumented.

Where the intermediate material actually is

Not in tutorials. Four places.

Issue trackers and pull request discussions, where real problems are worked out in public with dates attached.

Conference talks, which are frequently the best intermediate material available — somebody who solved a specific problem describing it for thirty minutes. Their limitation is the polish, and even polished they beat a generic tutorial.

Documentation of the specific tool, read properly rather than searched. Most people have never read the whole of a tool's documentation and most of it is short.

And other people's code, which is the least comfortable and most reliable.

What follows for anyone writing

Date it, and name the versions. The single cheapest thing that makes technical writing survivable, and most material omits it.

Write the specific thing you just solved, not a general guide. Specific dates better and helps more.

Include what did not work. It is the part nobody else writes.

And accept the small audience. An article helping forty people with a real problem is worth more than one helping four thousand with a problem already covered a thousand times.

The short version