Hacker News Reader

Comment by macmac | Hacker News Reader

Could provide actual arguments for why these aspects of Datastar are "crazy", "even more crazy" and "wrong"?

What more do you need than writing something like `data-on-input__debounce.200ms="@get('/examples/active_search/search')"` in the DOM as being crazy and wrong?

To me, it doesn’t seem any more far-fetched than, say, instantiating a virtual DOM in the browser.

So your answers is no, you cannot provide any arguments.

I think OPs point argues for itself. The example provided is awful developer ergonomics.

Okay but why. Surely if you have auto complete this works just fine.

The poster provided an example of poor abstraction. HTML for display, JS for client side logic. Combining them is part of the problem with other popular frameworks.

No, it's the entire point of datastar - that you don't need much JavaScript at all because some simple html attributes do most of it for you, and any extra can either go in-line, or you can also just fallback to separate js scripts if needed.

I think you, and others, just don't understand what hypermedia is or what the idea is trying to achieve.

Which is weird, because this is (was?) the predominant way to make websites for decades.

You mean like onclick?

I mean, that seems ugly to me, but... Makes intuitive sense and is fine actually? Is HTML a rich language in which to embed stuff like this? No. Does this clearly get the point across? Yes.

[Edit: this should've actually been attached to the GP comment. I agree with the parent.]

Isn't that the whole point of hypermedia frameworks like Datastar and HTMX? It's to extend the declarativeness of HTML.

If you don't like the syntax, that's your preference. But I'm curious to hear why this is "wrong". Since that's a more objective thing, we can discuss it

Because it's a mish-mash of different ad-hoc DSLs.

You have a DSL based on data-* attributes

But then some attributes get this weird syntax-ish extension

Then some of those attributes accept a Datastar expressions DSL: https://data-star.dev/guide/datastar_expressions

So you can have

   data-on-click__debounce.200ms="$landingGearRetracted && @post('/launch')"

Which is just bad design

I can see what you're saying.

But again, the API design woes are subjective and I think it's perfectly fine to iterate on getting this to a nicer state.

"Wrong" to me suggests a gap in the understanding of fundamentals or of how things work. If the ideas of Datastar are fine to all of us, and our issues with it are ergnomics, then that can be a more focused discussion.

> But again, the API design woes are subjective

They aren't. This is not good API design

    <button data-on-click__window__debounce.500ms.leading="$foo = ''"></button>

> "Wrong" to me suggests a gap in the understanding of fundamentals or of how things work.

Snide vaguely dismissive remarks don't make such remarks true.

> If the ideas of Datastar are fine to all of us

Who's us? Are these "us" in the room with us right now?

We are limited to what is possible in the data dash-* spec. If you have better syntax that's going to work everywhere in every browser please let us know!

You literally have a custom Javascript-like DSL already https://data-star.dev/guide/datastar_expressions

This wilful insanity is completely incomprehensible to me (HTMX and lit are also fully infected with it): "Oh no, we are just HTML, we can't do anything" while literally doing tons of things outside of HTML.

What's your alternative that stay HTML spec compliant? Instead of whining show code

> What's your alternative that stay HTML spec compliant?

Again: you literally have a custom Javascript-like DSL in Datastar. Use that.

It's like the "it's just HTML" or "it's HTML-compliant" mantra somehow damages the brain, or something.

Edit: this custom JS-like DSL is so prominent and such a crucial part of Datastar, that it's referenced in the very first paragraph of reference: https://data-star.dev/reference

Oh, look, your HTML-spec-compliant thing in which it is apparently impossible to do anything outside data-* attributes somehow calls external functions, and updates signals, and reads signals, and does all sorts of things:

    <div data-signals-result>
        <input data-bind-foo 
           data-on-input="myfunction(el, $foo)"
           data-on-mycustomevent__window="$result = evt.detail.value"
        >
        <span data-text="$result"></span>
    </div>

But no! It's absolutely impossible to do something about data-on-click__window__debounce.500ms.leading

> They aren't. This is not good API design

Why isn't it a good API design?

> Snide vaguely dismissive remarks don't make such remarks true.

Agreed! I don't believe I was doing anything of the sort.

> Who's us?

People on this thread.

What design do you think would be better?

Both Datastar and HTMX have the same issue: they want to pretend to be HTML. So they force themselves into writing several DSLs like this to pretend that this is still "just HTML".

Since they have a full "Datastar expression" language, I'd just expand that. Then you wouldn't need these weird constructs:

   <button data-on-click__window__debounce.500ms.leading="$foo = ''"></button>

Well they are both "just HTML", Datastar being better about being spec compliant using the dataset attributes and all.

By expanding the "Datastar expression" language, do you mean a compiled language or moving all the expressions into JS files or something else?

> By expanding the "Datastar expression" language, do you mean a compiled language or moving all the expressions into JS files or something else?

Doesn't matter. They literally already have a Javascript-like DSL: https://data-star.dev/guide/datastar_expressions

Datastar is spec compliant HTML. Please point me to the code that's not HTML spec compliant

Reply on news.ycombinator.com