Live data from Hacker News

Apple proposes adding "model" element to HTML that displays 3D content

github.com

231–240 of 241 posts

Re: Apple proposes adding "model" element to HTML that displays 3D content

#231

No need for such an element. It's such a private use case to display 3D models that it's better suited as a web component than an element in the standard. See for example https://modelviewer.dev/

This should be much further up than where I found it!

As others have mentioned, the privacy implications are much more important. That is, the site owner will not have to have access to eg. the camera stream, but still use AR with this method.

So in a way, the value in this proposal is that it is less capable than the from scratch approach, and can be implemented safely just in the browser.

Re: Apple proposes adding "model" element to HTML that displays 3D content

#232
post #128

It would be a great idea if WebGPU wasn't a thing. Imo Apple and big corp should double down on their WebGPU support instead of pitching their own ideas. And if they want to support something that is production ready, WebGL2 has been out there for a while.

I believe Apple originally pitched the proposal for WebGPU?

Re: Apple proposes adding "model" element to HTML that displays 3D content

#235

Earlier quoted context omitted.

Sure, but that doesn't make standardization useless so people stop reinventing this same wheel over and over again for very basic needs.

I wouldn't call "display a chair in a web page in an inexact and implementation-defined manner, with air quotes 'interactivity'" a "very basic need". This seems like a very specific use case, which I think is best left to custom code.

Sorry "basic needs" refered to "basic needs" in terms of 3D rendering. It was an answer to people elsewhere in the thread complaining that this new element would not support lighting and other sophisticated 3D rendering features.

I agree that 3D/VR is definitely not a basic need for the web and i personally don't want/intend to use it. That being said, if there is an established 3D model standard, i think it's better overall if the browser implements it properly, rather than having two dozen half-baked JS implementations eating all your CPU/RAM just to display a tiny square.

Just like technically you can parse an MP4 video frame by frame using JS to display each frame in a canvas, doesn't mean you should. element is much better suited to that, and if that element had not been standardized, we would arguably have to choose between "Youtube browser" and Flash player fallback, which is arguably worse for everyone involved.

PS: Also as other have pointed out in the thread 3D on the web is over two decades old yet has never been standardized. See also https://en.wikipedia.org/wiki/VRML

Re: Apple proposes adding "model" element to HTML that displays 3D content

#237

Earlier quoted context omitted.

So true. I'd be curious about developing modern web standards without JavaScript. I'm aware of gopher/gemini but somehow i like the semantic aspects of the web.

There is no "semantic aspect" of the web. The semantic web is dead since invention. It makes no difference that you now call some boxes with some new names as those boxes get custom styling and behavior anyway. There is no semantics (no meaning) behind the names. Even there are proposals how to use elements "semantically correct" nothing enforces it. So it's useless in the big picture. You can't rely on any semantic…

> As a trivial example, everybody is free to put a h1 below a h6. It will "work"… Because there is just no semantic meaning to those names!

Well you took the most obvious counter-example, but there's plenty of semantics in HTML. See for example picture/source element, or audio/video. Both as a user and website operator who deals with people in several languages, i cannot have strong enough words for my appreciation of adding subtitles to a video using just `` without having to modify the media container server-side as previously.

Little features like this may sound insignificant to you if you only use the web for forums and blogposts, but they're in my view a great accessibility/internationalization win. h1-h6, unless used in a serious publication, has indeed no semantic meaning other than "make that one bigger than this one". But for deeper semantics, there's schema.org microdata and indieweb.org microformats, which enable deep interoperability/federation between websites with entirely different tech stacks. I would personally appreciate if either/both were standardized as part of the Web instead of the entire JavaScript craze, but saying the web has no semantics is ignoring decades of innovation in this field with very practical applications and i don't think that's very fair.

> separates again documents and applications(...). Gemini tries to get the document part right I think.

I strongly appreciate this separation, as i'm personally interested in document browsers, not applications platforms (i've already got an OS and desktop environment for that). However, i'm not sure gemini gets it "right" (very subjective). I like that there is a document format focused on plaintext UX, and i very much intend to support it on my websites sooner or later. I'm part of the tildeverse where gemini grew and is nowadays thriving. However, in order to reduce bloat, gemini has pretty much given up on every useful semantic aspect of the web. What's left in gemtext is: one-level lists, three-level headings, and preformatted text (code/quotes).

No more alternative text for linked images (unless you want it displayed for everyone including clients like lagrange which support inline images), no more federated mentions/replies like indieweb.org, no more scraping...

> Only in a model where you don't strive for pixel perfection and customizable behaviors for every piece of rendered content something like semantic meaning of those peaces could arise (...) this model is not the web. Not any more…

I'm not sure that's entirely true. Some part of me agrees with the argument, but on the other hand i don't see a reason why a proper semantic markup couldn't provide a suggested pixel-perfect layout which the browser could freely ignore to make its own. Also, when you disable JavaScript entirely, the web is a very capable semantic document platform.

All in all, i don't have a clear opinion on the topic. But i'm sure it's worth investigating. I'm all ears for resources and projects in such directions (either more semantics in gemini, or less interactive stuff in the web). Thanks for taking the time to come up with a thoughtful answer!

Re: Apple proposes adding "model" element to HTML that displays 3D content

#238
This is hilarious considering I distinctly remember making some webgl games as a part of a project at a digital agency in like 2015 or so and Safari was the only browser that didn't support webgl unless the user specifically enabled it in the settings, making it a massive pita.

Re: Apple proposes adding "model" element to HTML that displays 3D content

#239

Earlier quoted context omitted.

The object tag is problematic because it has a single interface for every supported content type . Want to expose functionality? Can't, unless it's a param that is passed as a string. Having a dedicated element type enables it to be used in many ways (like drawing images to a canvas) that an object tag does not support. It's also hardly a case of NIH to update a spec with a more specific element than another one in t…

> The object tag is problematic because it has a single interface for every supported content type. Want to expose functionality? Can't, unless it's a param that is passed as a string. On a markup level, that's true of any tag. On a functional level, it’s not true of , as the tag expects supported functionality to be defined separately for every type. > Having a dedicated element type enables it to be used in many wa…

Yeah, if we all used object instead of image tags then you could just download an extension that filled in avif or webp and we could move forwards.

Re: Apple proposes adding "model" element to HTML that displays 3D content

#240
post #78

Earlier quoted context omitted.

Netflix make money from their subscription service not DRM. I can easily torrent anything I watch on netflix but I watch it on netflix because its easier. Thats why I pay for netflix its not because their DRM prevents me from pirating anything.

Sure, but do you not think Netflix would have issues acquiring content if users could right click->save everything they viewed?

Thats a good point the DRM is more for netflix's partners and netflix is nothing without them.
Post reply on HN