gvkhna

Pretty self explanatory, I’ve always wondered why jupyter notebooks today don’t support other languages like javascript. And would jupyter support things like mojo or Julia?

I specifically mean an opinionated all or highly encompassing web framework. Most modern web frameworks are far more modular, whether for node or even go. I understand separation of concerns and different needs. But the popularity of rails back in the day (and still today but “slowly”) to quickly get many pieces working was clearly an advantage for many startups.

Now I feel like the goalposts are always moving for newer languages, new libraries and communities every few months, constant migrations to the latest and greatest.

I’m wondering if it’s simply DHH opinions, ruby’s allowance for magic via method missing, or simply too much focus on performance, that modern web frameworks are not quick to setup and use. By web framework I don’t just mean database and views, but also, auth, storage, image variants, job queues, development and production environments, admin backend, caching, good migrations. The list really goes on.

Yes you can use 10 different saas co’s and patch everything together, including things like cloudinary, supabase etc. Is this really better?

Or the only way to get way way faster is to simply abandon rails but take on a lot more development burden?

Just out curiosity I’m wondering if HN has a decent size self hosted audience. Do you self host?

I self host many services on an Unraid, and may eventually move to TrueNAS Scale. It’s not something I spend too much time on, but I rarely hear about Unraid or Truenas on HN. Maybe it doesn’t make it to Front Page because it’s not very popular. Upvote if you Self host and tell me a bit about your setup.

The only reason I self host is control, I gave up switching SaaS tools every couple of years to the latest and greatest. I appreciate that a lot of FOSS still works great after 30 years, it’s boring, dependable, and slow to change. Mostly set and forget. Unraid not having a k8 cluster is starting to show its age, but TrueNAS Scale seems a bit too immature right now. Perhaps in 2-3 years it will be boring and stable. Share your thoughts!