Anything Web Component related is not worth anymore. Sadly. I prefer collaboration than hostile standardization. news.ycombinator.com/item?id=4708...
20.02.2026 12:49
👍 0
🔁 0
💬 1
📌 0
Anything Web Component related is not worth anymore. Sadly. I prefer collaboration than hostile standardization. news.ycombinator.com/item?id=4708...
@lit.dev is cancer, "is just html" when using their influence to carry to the browsers via specifications their own experiments, and making the research, a "Frameworks vs WC" kind of thing. Conflict of interest, rot from the inside, unfair competition. Im happy frameworks colaborate, not dictate.
If by participating, you mean the following, yeah, no wonder..
bsky.app/profile/pass...
bsky.app/profile/konn...
bsky.app/profile/just...
If it can refactor dom-expessions I would like the URL of your dealer 😂
If people do not want to move forward with input.value being an attribute, then so it be, give us tools/apis to handle this instead of trying to paint the world only with the colors you provide. Give us tools, so we can improve our mechanics instead of forbidding what we want.
Thanks to you all, I came to the conclusion that things should be an attribute, and not based on heuristics, but still, there are some expectations, and we should work with these expectations in mind.
The WC group doesnt listen. The world is not black and white (WC vs frameworks). Your opinions are too rigid, and you wont accept shades of grey. Such `isWebComponent`, React and Solid insist on `<input value=""/>` being a special case prop not an attribute.
I've been looking at `html` tagged template literals again and man the state of things sucks. I'd be so happy if @developit.dev `htm` was the accepted format and we could just get typing going, but as usual there are competing standards and in a lot of ways it makes sense.
Yeah, the defaults, having to ask people to switch a setting (embeddedLanguageFormatting) is not workable. They would lowercase any tag that matches native, say `<A/>`, which means we cannot capitalize any component, as being added to html later would mean that prettier will start lowercasing it.
Component registration, also allows registering components globally and locally. Say `html` that includes `For/Show` and friends, and local versions that exports components of some library
inline case sensitive, doesnt really goes against standards, one can pretend is some sort of xhtml, reasembling JSX. If an `<div class=""/>` tag can be typed at some point in a tagged template, then also we should be able to type `<Counter steps="3"/>` by extending something like JSX namespace.
It depends, generally, I would say the guide for users is in encouraging patterns/providing utilities that avoid the issues. At the same time, if it can be solved in runtime, consistently, without introducing other special cases, or being too fragile, then maybe even if complicated could work.
Sounds good, I'm curious what's the issue with it, or what demonstrates?
pd: FOUC is Flash of unstyled content, not "Frame". I suppose a Flash could happen within many frames. 🤓
This is cool to figure out where to Suspense, and also makes me think that single-flight request could resolve earlier even with a pending resource, unless the response is already being resolved separately. Fun stuff
Simple apis additions make me happy
removing 11k LOC should be everyones goal
The force is strong, html6:
- strict parser, no autofixing html, back to xml?
- self-closing tags
- fix boolean attributes
- remove useless tags
- simplify web components api for sharing components between frameworks/js
- simplify dom apis
That's like @trusktr (Joe Pea) htmlEffect that used to have an update function for when not using reactivity. That, for a single module may be doable, but then you tangle many modules, and you are overcalling update. The beauty of reactivity is on derivation. What maybe we are missing is forcing derivation a bit more, but there's some work about that on solid v2. The almost invisible performance cost of a reactive system pays off by the consistency provided by the state of the state. Manually syncing things is prune to error because well we aren't perfect, specially when working as a team.
Recycling a discord message on `this.update()`
En lo posible pasar una noche en Cabo Polonio y rezar para que no esté nublado, porque de noche mirar al cielo te vuela la cabeza. De paso una empanadita de mejillones.
Ramón Méndez Galain helped Uruguay decarbonize its grid in just five years, with 98 percent of its energy coming from renewable sources. He now aims to help at least 50 countries embark on the path to more clean energy.
adorable, so cool to see it become a real item! quack! quack!
For future js/html surveys, could be possible to consider to always include a question on the lines of "are you required to write javascript-less forms/widgets?", and ask to clarify why so. Example: gov websites/accessibility, etc. I'm interested to see what people could say. Thanks!
The way I see it
- events do not collapse (firing events may put you in the initial state [no optimization])
- events do not model derivation, so may lead to have to do manual synchronization of state, which causes glitches (main incentive for using signals)
@ryansolid.bsky.social mentions in an article, something on the lines that **mutation** behaves or relates to events.
State of HTML is live now!
Please take it if you care about the web platform! 🙂
These surveys are super helpful for our team to understand the state of the ecosystem and prioritize missing features/platform needs!
survey.devographics.com/en-US/survey...
the comments deletions mess up my posting btw
Kind of uncalled-for all these comments
@webreflection.bsky.social
In any case, I'm referring to the fact that if you eval third party on your service worker that it will behave as if you were able to load it from anywhere, which is what you complain about.
Seems to me, you misunderstood.
your post shows code is being loaded from a third party, esm.run, so what I said applies,whatever. btw I was looking in Mdn and it seems possible to load service workers from other domains with some headers, but not sure, took a quick look
Can't believe its 2025 and people still try to `eval` things
