[Python Dependency Pitfalls] A total mess?

Hey there,

Recently I watched a Pythonista ask for advice on setting up a Python project on his work machine.

This new developer had some prior experience with NodeJS and had just started to get his feet wet with Python.

But OH BOY was he already frustrated:

"Why is Python dependency management so sh*t?"

He was trying to install someone else's Python application and to get it to run on his work machine.

The app had some external dependencies it needed to work properly—things like a third-party networking library and a database interface.

But the way the original author "documented" these dependencies was utterly lacking:

There was only a small "README" file with a list of packages the project needs to work.

That's where the trouble started—

This list was a) incomplete and b) lacked any versioning information whatsoever.

So our poor Pythonista is trying hard to make this application work.

But every time he installs another package dependency and then tries to launch the app he gets slammed in the face with this:

"ImportError: No module named X"

He then spends 15 minutes googling for a solution, finally hunts down and installs the missing package, runs the app again and:

"ImportError: No module named Y"

Yuck!

(Rinse and repeat.)

He's spinning his wheels, and the lack of progress is starting to get to him.

After all, what he loves about programming is "programming"—and NOT wasting hours on trial and error just getting ready for some real development work.

So the guy becomes (understandably) more and more irritated.

Finally, he vents his frustration by putting the blame on Python:

"In NodeJS it's fairly basic to have all dependencies listed in package.json and everything gets installed automatically with a simple command…"

"Why isn't there such a thing for Python? Python packaging is A TOTAL MESS."

And who would disagree with him after going through the same experience?

Modern dependency management and dealing with third-party packages in Python is complex.

And it's easy to get stuck or to "shoot yourself in the foot" if you don't feel confident in the central workflows and tools.

Over the next 5 days, I'll show you some of the main pitfalls with dependency management in Python—and how to avoid them.

When you steer clear of these potholes, you'll find that you can make progress more easily…

And you'll pull ahead of your peers who are left wondering why their productivity continues to plateau.

We'll kick off the first lesson tomorrow.

— Dan Bader

P.S. Ironically, pitfall #1 is something that recently shook the NodeJS community to its core.

It was a "clustermug" so bad that even some newspapers reported about it. But unfortunately Python isn't immune against this problem either…

You'll hear the details in my next email.