No dev, just a dev rant.
Whenever I think I would like a little sandbox to do things in the first question that comes to mind is "What will bind it?"
Now I could just go with Laravel and not have any issues and live a happy life. Yet that can seem a bit like stagnation, specifically in the web application arena.
Since my first project is a simple P5.js L-system generator, and at this stage does not need anything fancy like data from an API or whatnot. I figured it would be a good call to have Sveltekit be the binding.
Routing in nodeland is a load of shit as it is, their concept of SSR is a rubengoldberg machine and you would think rendering a simple client-side html page with some libraries and two-way binding for some bits would not be rocket science at this point.
The two-way binding is not, fair enough. Everything else leading up to bring a client side page together in a SSR obsessed js framework... Is a bit of a shit show.
Now sure I don't know all the ins and outs and I can understand that in the fight against react infested and bloated js apps there will be some sacrifices to be made. Like sanity.
I stuck it through using Sveltekit and unwittingly seemed to have chosen a library which could be considered its arch nemesis in general.
The deal with p5.js is that it thinks it is the only library anyone will use, it even by default uses some fucked up script tag dumped onto a page... Needless to say I will be changing that out because even in instance mode it just does not play well. Isolating P5 to just deal with P5 things is horrid.
Regardless, Sveltekit still somehow found a way to just screw it all up when I finally built the thing and pushed it to production. Things I never even considered because DEV is nowhere near the correct output when in production which again makes no damn sense, but anyway.
It was all working quite well and I kept thinking ok I just need to get use to how Sveltekit does a couple of things... No, at a point it is not getting use to anything, it is broken and I have no need to deal with over-engineered bullshit.
Why I say P5.js happen to be the arch nemesis is because even after I figured out how to get it to be client only, Sveltekit went and built the project as one big lump of shriveled dick cheese. Then Vite the thing it uses under the hood goes and imports imagess etc using a logo = new URL() syntax and guess who happens to have a nice little function like that and did not like Sveltes little logo import one bit? Yup, P5.
So needless to say, things just don't mind their own business. The lack of seperation of concern if that is even the right phrase is infuriating.
Now I could also go with NUXT but really there is too much fluf for what I want to do and I found state management in NUXT to be a pain in my ass. Generally I just find NUXT a bit confused.
Obviously there is no way I am touching any inbred REACT system since for all their bragging I am unaware of any benefit they have brought to the JS ecosystem ever.
I am looking at ATOM though, I remember looking at it before and the benefit would be I get to integrate Svelte components, Alpine, or Vue ones into it quite simply it seems.
So that already screams "binding" , as mentioned I could just go Laravel or heck plain html and just link to pages that contain these components. There is always a backend bit to things and some backend bits are nicer in node.
Now if my binding framework can handle datastores, db and workers that go about doing stuffs then that is what I want it for, then each page or section I can play with different libraries to use said data and or shared functionality to build ontop of however I like.
If like NUXT or SvelteKit they are too library centric then it all seems to fall apart. So I guess my mistake was using SvelteKit thinking it is a true backend framework aside form just a glorified router for Svelte components.
The same goes for NUXT it may well be backend but it has way too much boilerplate and fluff, although it is far better at it than SvelteKit.
I might just decide against Astro but from a quick gander in my rage against SvelteKit I will stick to it for now. Since I have alrady built the P5 version inside SvelteKit it will be fun to see how easily I integrate that svelte component into an Astro page.