[ivanb]

The ultimate failing of Clojure as a tool is its inefficiency measured in $ per feature.

The language is not very approachable for junior programmers. You can grab a Python programmer and graft them on a Java or TS project and the same week they will deliver features and bugfixes. Reading other's people code in Clojure is much harder especially when coming from imperative world. The language is terse. The documentation is terse and lacks examples. Library documentation is insufficient or non-existent. You have to read other people's code. It is a really demanding activity aggravated by lack of static typing. You cannot just hover over a map in your IDE and see its structure. The value may have come through several transformations that add, transform and remove fields. To understand what arguments a function expects you have to decipher destructuring bindings which in real code can be ingenious, i.e. unreadable. And hold on to your butts when a clever teammate invents a flow of control macro and uses it throughout.

These and other reasons make the pool of candidates small, candidates "over-qualified" and expensive. If your expert quits, your whole endeavor is screwed.

I helped to rewrite a Clojure project that stalled due to lack of affordable talent. It's been chugging along since then at a considerable pace. The new language is very popular and much less affected by the mentioned problems.

There is a mention of Penpot in this thread. Mark my words, when it comes a time to scale, they will rewrite the whole thing in a popular statically typed language.

[mnming]

I do agree. While I made all my projects in mostly Clojure, I absolutely cannot imagine hiring & onboarding other drvs.

And if I had a tough day, my clj code will look like rubbish, while my typescript code will still look decent.

It's really not a language for big team.

[simongray]

I have onboarded developers coming directly from university and it went fine?

In fact, at my old job, the entire team learned Clojure on the job. I was the only hire who had learned it prior to being hired. Our codebase was fine and very maintainable.

The other teams were similarly new to Clojure. Most people spent a few months doing increasingly more advanced things until they were comfortable with the language and the codebase.

The main issue for complete beginners is learning to grapple with the different development setup compared to what they might be used to, i.e. interactive development on a live system using an editor-integrated REPL. We basically set everyone up with IntelliJ and parinfer. Paredit was optional, as was alternative editors such as emacs.

IntelliJ + parinfer is a similar development experience to Python. The only part that is foreign in that case is the interactive development of a live system which the new devs quickly come around to seeing the benefits of. It helps having seniors who point out the new concepts.

[ballpark]

> It's really not a language for big team.

Not my experience, but you do need senior devs who are good at Clojure to be reviewing and helping the juniors

> if I had a tough day, my clj code will look like rubbish

This may not be a reflection of the language

[vfclists]

Clojure is not Go or Python