Ask HN: What's your biggest issue with HTML?

HTML is not fit for some of what we currently do with it. Its print roots are obvious and _relatively_ well adapted for structuring editorial content. There are still many maddening issues when you come from a print background but it is not always clear if they are on the spec author or on the implementor: bad justifying and hyphenation, vertical displacement caused by sup and sub, etc.

But the big issue is all the UI stuff that goes around that editorial content, that we have to reinvent all the time or use piles of third-party stuff for. HTML has no tabs, no cards, no popins, no mega menus, no button groups, no steppers, etc.

We shouldn't use that "platform" to build applications.

...that it still exists.

God, I wish we just had a better language for the Web altogether. Something with sane UI primitives that automatically translate into platform API calls (i.e. you say <button>, the OS renders a button based on its own UI guidelines).

Some days I'm still nostalgic for XML + XSLT, which is like a super strict but super powerful alternative to HTML + CSS. You define the structure however you like, using a schema of your choosing, and then use XSLT to define the output, transformations, and styling. Instead, we get kludges like React trying to shoehorn their own abstractions into a million autogenerated nested <div>s because the DOM is the DOM, and having to build a UI lib on top of that, once per project per company per platform. It sucks!

HTML is a blank piece of paper force-fed into a 90s-style typewriter. Modern UIs are composed out of layers of abstractions that vaguely signal intent. The two are not a good pairing :(

I don't have any gripe with it other than one thing:

Lack of widgets. We have people writing emoji pickers in JS and placing them in documents, and some of them work horribly on mobile. Then there's the caveat of emoji pickers being already available on mobile and no such JS widget is needed. Then there's 100+ variations of dropdown menus implemented in JS, which all have their own bugs, don't work well on mobile, and are already somewhat natively available on mobile anyway so the JS widget is not needed.

That it's locked at where it was over twenty years ago (modulo Ian Hickson's additions ie article, header, section, etc) with everything around it (CSS, JS) having been bloated ad absurdum to make up for HTML's lack of evolution. Yes I know we want to call it separation of presentation from ... whatever, but looks more like an organizational stalemate and self-inflicted problem to me. Today, WHATWG HTML the markup language specifies a vocabulary for casual academic publishing and in prose rather than formal fashion; when in reality it's used for messages and discussion threads, ads of all kinds, CVs, press articles, forms workflows, and countless other kinds of text. To capture such uses was the goal of W3C's XML initiative in 1997 for bringing new vocabularies to the web; but they obsessed over "meta" and enterprise techniques and only brought SVG and MathML. Subsetting XML from SGML was an entirely pointless 20+ years detour, and today SGML still remains the only game in town for formally capturing and evolving HTML. Hickson's/WHATWG's HTML spec OTOH is already fossilized and error-prone to update with new elements.

Some redundancy in tags (<b>, <strong>), forgiveness when it comes to coding (you can type anything, no errors, no problems, every browser decides how to interpret the code)

Most of HTMLs problems were solved 20 years ago with things like schemas, namespaces, and conventions for extensions supported in browser. None of this caught on. At the time people wanted sloppy shit because all the value was in business logic written in ASP or Java that wrote static HTML with dynamic content for advertising and e-commerce. The just had to barely work well enough to create an e-commerce conversion and avoid lawsuits. You have the same shitty code now but it’s moved to Angular and React.

When people claim they want HTML improved I stop taking them seriously.