Ryan Carniato

I don't want to take up much more of your time. I do recognize the futileness of trying to change the past. This whole discussion didn't start because I was mad or annoyed with WC folks, I was just commenting on the state of things. Which I understand. People take that as being angry. Its fact.

My oldest open source library was to allow creating Web Components abstracting away the element part (I wrote this 2 years before I published it openly using it at my company). The element part was actually the most annoying part of it.

www.npmjs.com/package/comp...

I chalk this up to bad timing. We moved too fast on WCs at a time when we were only just starting to understand what components on the web should be. I don't know why we did. I get the need. I really dislike the solution. I realize this is probably a really long held bias too though.

If I had to go here, then DOM elements have good and bad properties for the role. I can think of different solutions that would align with everyone better with what I know now, .. even maybe what I knew 10 years ago. I feel around the time framework authors realized ~2016 might have been the time.

This situation is nasty regardless. I see how things get pushed down a layer. Nothing I'm describing is actually desirable and while I could argue we are solving the wrong problem, that wouldn't be a popular position to take. Even if I have concerns about our design system passing off as native.

You kinda invited it. We just increased the number of islands. Size/performance aside, we distributed the maintenance burden. But presumably we only have one at a time. There might be versions of those WC if attrs change but you could probably keep things backcompatible.

Realistically those components will be updated and replaced many times over time. They will probably be written with different libraries depending on complexity. you might have every component written with a different one/version. You didn't really solve that part but it isn't your intention to.

Yeah it's the interface that is the problem. There are benefits to using an existing one. Although the result of which actually changed the identity of it. But I think we need to look at separating the mechanical need from the solution.

It's being able sit here and be like maybe they weren't a good idea. The challenge here is here is some expectation of what instead. In some cases its like "nothing". But I think it would be interesting to revisit the fundamental needs. And not be like we don't need Scoped CSS we have Shadow DOM.

We can't clean slate the implementation. WCs are here to stay. But maybe we can clean slate the thinking. Make them an afterthought. Part of filling out the picture not driving them. Most of us have adapted to whats there (maybe not React).

This is why we don't generally participate. You present it so that you leave the door open. Appreciated. But like it isn't going to change our mind on this. I hope this can impact future decisions. But I get why that would be difficult.

I think most FW authors are in the same boat. The thing that ticks us off isn't that there irreconcilable differences of opinion, it is that you think your airplanes are somehow different. I mean they are different (in ways that we think matter). But they are still flying devices.

You aren't being rude. My position is and has been for a long time that some things are like gravity. I appreciate you all building airplanes. I build my own too. But there is some undeniable physics here that I can't ignore. I never claimed I'd come around on this. I am willing to talk.

I get WCs were built by committee so like no party is solely responsible and we got here together. A handful of framework authors seemingly pointing fingers isn't helpful. But engaging them on this isn't unless we can start from a conceptual place that WCs were a mistake and look at alternatives.

It is much more tempting to continue to show through building what is actually necessary versus not. But that won't curb people trying to build standards to solve their user space problems. I mean that is why they are there, but man the cost.

Just hypothetically what is the best course here? It's hard to take back what is already there. Im' sure people would like what is your alternative? What if the alternative in many areas is there shouldn't be a standard in that position. Nothing to replace it? How will that go?

Worse the source of this divergence continues to poison the well. Lets say I do want to get involved I do want a better web. I do want to put in the time. How do I enact real change? Do I show up at all the meetings and fight every related proposal, to try to course correct?

Look at it from this perspective. I build systems. I'm very particular on solutions otherwise I wouldn't do it. I care a lot. If I know that while I could make something work it would never be the best choice, not due to a today limitation, like categorically, systematically, how do I reconcile it?

There are shared goals. Motivations. Inspirations. But yeah it is hard when the fundamental implementation concept is at odds. Everything stems from that. Truth be told the damage is already done. It is too late. So why not be part of the solution? Because it's the wrong one. It's that simple.

I have these discussions with the React devs too. The difference is we recognize we have a different perspective and don't assume we can change the other persons mind. We understand those differences. I've grown to respect their position. But they also don't threaten the foundations I build on.

Yes. That is hard. I'm open... (I hate saying this because I've done this so many times) to discuss it more if you would like. My perspective is purely technical and argumentative. I don't care about success. I never did. That isn't why I do things. My points aren't obscure they are systematic.

I guess.. I feel like I've done this 3, 4, 10 times already. Maybe it is a me problem, but I've spent the time each time writing my thoughts. I get why there is a disconnect, my perspective drastically different. The best this accomplished was people acknowedging I have a point but its too late now.

I mean what do you want to pin me down on? I think I'm being pretty clear on my position and justifying that.

I understand the desire to find this middleground. But the challenge here is while I can see how people find value here, it doesn't change my perspective on the facts. My view is very technology driven. I'm definitely de-prioritizing the human aspect. But I'm not unsympathetic to the motivation.

I wouldn't keep answering if I wasn't trying to have the conversation. I think the fundamental view point is just so different it's a difficult conversation to have. It's hard to be constructive about something that you believe is fundamentally flawed. If you want something concrete, it's that.

Maybe there was ambiguity before. But for some of us elements never should have been custom. They were the building blocks we could rely on. HTML is an output format. DOM an manifestation of that. Making it anymore has considerable consequences.

The irony that there has been push back there. The only inevitable part of all of this. We are here now. But I take this to be a great lesson in how easily if we aren't careful we can change the fundamental identity of a thing through extension.

It's hard to expose lower primitives when concerns handled by CSS, JS, and HTML become so entwined. Like building on top is easy, JS or HTML both are great carriers. Your component could be a script tag. Neither are guarenteed to execute in other envs. But changing built ins is hard.

Inevitability is an interesting idea. I don't believe CE are inevitable from a modular sense. We don't need our modularity to be based on a dom element. Components conceptually will always appear but they could take many forms. The actual need was probably customization of native built-ins.

If it has reduced down to a small range of specific decision points that need to be made. Maybe. But if every framework has a different version of the thing and would want to beyond historical reasons, not a good a candidate. I think that misses the possibility of things that could be, but a start.