@iamdtms I used to be a fan of the semantic web; unfortunately, that idea died hard sometimes in the late 2000s.

Nowadays, I wish we would abandon the entire web stack and instead share data we can use and process locally.

As such, we are lucky that tools like git were invented before everything became a webapp. Imagine if code revision systems were invented today. Everyone would be forced to manipulate code in a browser. The very thought makes me sad.

Running: I haven't tried the brand yet. It is not widely available in Denmark, but I did manage to find a store that sells it, and I received two bags yesterday. I have been using another similar product, and when I use that for long runs, I can keep going for longer and have a productive day afterwards, which is good. I don't know how healthy it is in the long term.

CSS: It is too granular, and DaisyUI is the right way. Otherwise, it feels like writing assembly code in the class attribute.

@GeeBee build tools because the web stack has always been a bunch of hacks strung together with duct tape; moving fast and creating huge exploits and patching these security holes has been the name of the game by ruthless browser vendors eying total domination of this “open, write once, run everywhere ;)” platform.

Human beings should not write this stuff by hand. Hence, the build tools. If the web should be, it should be a compile target.

(I used to be a happy frontend dev)

@GeeBee initially abusing the class attribute for these micro, CSS-centric classnames feels wrong. But I think it solves a huge problem. Because HTML is dynamic, and a class name can come and go, even as the page is loaded, there is no way to write a dead code eliminator for CSS files. If the css is defined in the class names, it comes with dead code elimination.

But the “widgets” defining the class names needs to be hidden and abstracted away—because it is ugly and unreadable