Plain – a web framework for building products with Python

Related, "Django's REST (Framework) Problem" — https://news.ycombinator.com/item?id=43510495

I'm not sure that many people who rely on Django Rest Framework are aware that last month the bug tracker was made private and the project is looking for new maintainers.

I love Django but the project needs to go through something similar to Angular's renaissance (and Angular needs to learn from Django docs.) I'd love to help but it seems that most of the efforts to address the issue have been stalled in committee.

A fork probably isn't the answer but something needs to be done. If it's a money issue, pass the plate! Whenever I talk to Django devs about contributing the feeling that I'm left with is that I could put in years of work, jump through every hoop, and at the end of it they may still say "We're not sure."

The feeling that I've gotten is that the Django dev community is very small and tight-knit. Whenever I've talked about helping out on various projects I've walked away with the feeling that their friend is handling it and they'd rather leave them to it. The community has been trained, through years of reinforcement, to wait instead of getting involved.

Jacob Kaplan-Moss gave a great talk at the last DjangoCon about money - what Django would do with an increased budget and how to get there: https://jacobian.org/2024/oct/8/dsf-one-million/

Video: https://www.youtube.com/watch?v=5nS1SSuHk9I

I saw the talk last year. Let's do it.

If I had $100,000 to spare I would give it to Django as an unrestricted donation. It would be worth every penny. But I'm a solo-dev and don't have that kind of cash on me.

Here's what I can do. I will send you $1,000 today if you can get something like a `django.contrib.rest` package (i.e. official rest api support) on the roadmap and secure matching funds. I'll make it a recurring donation if development gets going.

I don't want to support other REST libraries because there's no consensus. One app ships DRF views, another Ninja, and that one over there still has a Tastypie dep. Remember Piston? Using multiple rest packages means that figuring out how to create a consistent error format, etc is an exercise left to the reader.

Django makes great design decisions that the majority of devs can get behind. I'd love to have an officially supported REST package that feels like the rest of Django.

`django-tasks` is a good recent example. I was reading through the code a few months ago and I was impressed.

Another project I'd fund: Making Django an async-first (not async-only) framework over the next decade. It's easier to mix sync code into async than the other way around.

I just setup a $83/month ~$1,000/year recurring donation.

https://www.djangoproject.com/fundraising/ if anyone else depends on Django.

thanks for sharing this, just donated too.

I read thru Plain's doc, it is not bad at all, but I really hope Django can be much better than it is now.

Yep – I've been using Django since 2007. The big win used to be the admin, ORM, database migrations... but now oddly enough a lot of that has become a pain. I'm someone who knocks small solutions together for fun or to scratch and itch, so I'm looking for low maintenance. The problem I need solved has shifted and now Django is too much boilerplate (APIs and models are perhaps too distant as concepts), and too much maintenance work. Auth is perhaps underemphasised as an area for improvement. The built in auth isn't really fit for purpose anymore, and the various extensions for federation / passkeys take work to integrate and change a lot.

None of this is to write off Django or the people who've worked on it: I'm genuinely grateful for the framework. It's let me build open source things that help people out. The typical problems most of us standing up small-to-medium solutions need solved by a backend have just shifted underneath the framework, and it hasn't had the resourcing to keep up.

I've been looking at Pocketbase as a replacement. I think I'd prefer something that uses Postgres rather than sqlite, but it's pretty awesome as a solution for those two or three day projects, and the maintenance burden looks like it's pretty low on an ongoing basis.

So glad that's not actually true so I didn't have to rewrite sqlalchemy from scratch

It's more straightforward to call a sync function from an async function than the other way around. That's a pretty uncontroversial statement. Adding async support to a library doesn't mean starting starting over.

I'm not fully understanding your thought. Can you say more?

Django dying the Drpal death of becoming a generalised case of nothing but itself?

Try Flask.

If one can make it without breaking the API too much, why not.

The #1 reason I use Django is stability. I have projects that span 4 major Django version without any significant break. That's quite a feature.

What's become a pain about DB migrations? They've barely changed and they are still so amazingly useful that you forget it's something you have to think about until you move to another framework that doesn't have them.