Kilian Valkhof
I don't know about you, but it feels like things are happening in the base layer of the web.
New Capabilities
Browsers are trialing new elements that let users grant permissions (e.g. for using the camera) without requiring JavaScript. We're so close to finally getting customizable select in all browsers, and other new form capabilities are cropping up: color, switch, showPicker() and perhaps even a new rangegroup element. And features like <dialog>, <details>, and the popover attribute are all making all our lives easier right now.
What makes a lot of these APIs so great to use is their declarative nature. Why make yourself responsible for a click handler when command="show-modal" commandfor="dialog" is much less code, and handles so much more? Looking at the survey results, many of us agree. Overwhelmingly the content and HTML element-related APIs get the most love.
Something else that's become apparent this year is that the “base layer” of the web doesn't have to mean the user-facing web. Features like speculation rules (declarative JSON), declarative WebMCP and even the loading and fetchpriority attributes can help the user agent make better decisions around data loading and interactions, and if done well, the user shouldn't even notice them.
Lingering Frustrations
On the other hand, a bunch of features just don't seem like they're gaining that much traction with developers. Take datalist. The fancy suggestion box isn't something you can depend on even if it feels like using it gives you "select with a search box". And things like WebUSB or SpeechRecognition are fun APIs, but their use cases are limited. And what about <model>? A whole element just to show static 3D model files?
And of course, browser differences remain an evergreen source of frustration, but it sure looks like the problem has shifted from browsers doing things differently to browsers simply not having caught up yet.
This year in particular it seems that how mobile OSes and their UI choices impact the web is the main sore spot. Mobile operating systems are comparatively immature: will we look back in another decade and consider this period to be similar to the early days of Firefox vs. IE?
The Accessibility Story
It's also very heartening to see that accessibility is a top concern for survey respondents. And for good reason. You can't have good accessibility without good HTML, or good HTML without accessibility. They're not the same, but they're intertwined.
Accessibility is multi-faceted and that makes it hard for developers to know if they're doing the right thing for all users–and browsers have a role to play here to make figuring this out easier.
Editor's note: check out Kilian's very own Polypane for a browser that does just that!
What's Next?
What is left on our actual wishlist? After customizable select finally lands, we'll want to customize everything form-related (let's agree to start with datalist, yeah?).
What about HTML includes? Over at OpenUI, work is underway on things like tabs and focusgroup. Are app-like components now standard enough to become, well, standardized? Or will we shift to wanting HTML to take on more non-user-facing capabilities? Time will tell.
Looking back at what Dr. Lea Verou wrote last year, it seems that even though there haven't been huge leaps, we're definitely moving in the right direction as far as us devs are concerned.
Major pain points are simply gone with the broad support for dialog and popover, and customizable select will do the same for many a UI. HTML is the base layer of the web, so maybe it's not a bad thing to have it move more slowly and more deliberately. Our industry might be in turmoil, but I think our foundation is in good shape.