Ask HN: What language/framework would you choose to build a startup in?

4 points by RcouF1uZ4gsC ↗ HN
There is a post on the front page about why Rust is not a great language to build a startup in. All things being equal (not familiar with one language over another), what language would you choose to build a startup in?

6 comments

[ 5.5 ms ] story [ 26.9 ms ] thread
The one I am most familiar with.

For me as a potentially "technical founder", that happens to be Python, but obviously "I" is the important context.

All things being equal, I'd use some combination of:

1. Groovy

2. Java

3. Grails

4. Spring Boot

5. Postgresql

for backend stuff. For front-end, I'm still waffling a bit on what framework (if any) I want to really commit to investing time and energy in. I've dabbled with React a bit, but haven't fully committed to that as a strategic choice yet.

All of that said, on a recent thread similar to this, sombody brought up a couple of technologies that look really interesting to me, that I want to explore for possible use:

1. Hotwire

2. HTMLX

3. Elixir + Phoenix w/ Liveview

Arc / Racket / Common Lisp / Quicklisp

The ChatGPT method of language construction looks very cons.

Common Lisp does math right, which ChatGPT has to learn todo.

It depends on the problem you are trying to solve with your startup. That being said if you just want to build web apps quickly to validate your idea, Ruby on Rails or Laravel are two productive frameworks that you might consider
It obviously depends on what the startup does!

My favorite startup (2003) was building a shared-nothing distributed system. We chose Java for the distributed part. Messages were formatted in one of the language-independent serialization formats (Avro? Google Protocol Buffers? I don't remember). We wrote a number of distributed algorithms ourselves, e.g. leader election. We wrote our own logic to detect failed nodes, network partitions, etc., and to react to those events. It was a huge amount of effort, and took a long time to get right. Finally, it worked really well.

Shortly after completing the system I discovered Erlang, which had battle-tested solutions to many of the problems we solved on our own. That language would have saved us a LOT of time and effort.

Speaking of Erlang ... it's not just what the startup does, also where it is.

i.e. In a University town I'd pick Elixir/Phoenix ... great tech, but no dev pool ... but if your strength is abundant smart juniors, you offset the main downside.