7 comments

[ 0.21 ms ] story [ 23.5 ms ] thread
This is everything that's wrong with web development frameworks in the last 5-10 years. Everything's optimized for the seductive "get started in 5 minutes", one file example, and people spend the next 5 years trying to work around the framework's short-sightedness and come up with sane ways of organizing an ever growing application.

I sorely _sorely_ miss opinionated frameworks.

There are plenty of opinionated Python frameworks, you just seem to dislike ones that aren't.

This is also showcasing the ease of setup of Neon, mostly. Previous announcement here: https://news.ycombinator.com/item?id=31536827

Flask because it's literally anything that serves an API easily.

There is something a little cheap about these kinds of blog posts, I certainly agree, but you seem to have a lot of pent-up anger for it.

author here - guilty as charged re 'cheap'. I spent a few minutes on this to try out Neon and was pleasantly surprised that it did what it said on the tin so I took some screenshots, and then figured it might be interesting to HN.
It was interesting and the point is Neon, so thanks!
Strongly disagree with the downvotes on this post.

The phenomenon you describe is real enough, but I’ve also seen teams lament their shortsightedness as they fight with the limitations of opinionated frameworks (cough… Django… cough). It’s extremely hard to choose the right tech. The best engineers I have worked with are just as good at minimizing the stakes of making the wrong decision as they are at making good ones.

So um, why this over the one click Postgres in AWS,GCP,Azure etc...

> Neon is a new serverless Postgres database

I'm curious about how it's serverless. Is it not running Postgres... on a server?

Generally serverless means that you don't have to worry about the servers, not that it's not running on a server, of course it is. In this case, neon offers pretty good scalability that goes hand in hand with the usual serverless behaviour of having multiple instances of your code running in parallel, something which postgres alone is not very good at handling because of the "one process per connection" functionality. You'd need pgbouncer or RDS proxy on top of your postgres instance.