Skip to main content
Meta Stack Exchange

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

Required fields*

15
  • 3
    Much better, yeah. I think your underlying point, which I understand to mean that the design should be tolerant of user-side customization to let people customize as meets their needs, is much clearer and easier to understand. Thanks for taking the time to reword Commented Oct 6, 2025 at 18:22
  • 5
    @Slate It's not just about being tolerant: it's that this has to be accomplished in a particular way, which requires going back to the drawing board on a lot of things that SE doesn't really want to question most of the time, and bringing in expertise that SE currently lacks. (From what I've seen of the developers' and designers' work, they'd basically just need to be pointed in the right direction and given permission – but I don't think there's anyone who can do both of those things, or this wouldn't still be a problem.) Commented Oct 6, 2025 at 18:51
  • 2
    @wizzwizz4 I'm curious, you say "in a particular way" - can you tell me more about what way this is? Commented Oct 6, 2025 at 18:52
  • 2
    @Slate Empirically? No. I believe that I've been consistently communicating that for years, but obviously it's been lost in translation. I described it as the Lost Arts in this comment, it's behind most of my criticisms about Stacks in the Chat redesign, and I could probably teach it to a developer given a chat room and a week, but it's one of those things that I can't really describe except by repeating the word "semantic", and giving examples of what the correct approach achieves. Commented Oct 6, 2025 at 18:54
  • 4
    At its core, the idea is that HTML is for declarative descriptions of the content model, CSS is for declarative descriptions of the design philosophy, and JavaScript is for declarative descriptions of process and APIs, or respectfully shimming functionality that's not otherwise available. (Some of those "declarative descriptions of the design philosophy" ideas have made it into the current iteration of Stacks – some of the named CSS variables, for instance, although others completely defy it.) Commented Oct 6, 2025 at 19:00
  • 2
    Part of being declarative, using a common language, is that where a browser API exists, and that browser API follows The Way (unreliable rule of thumb: if there's a Google press release about it, it does not follow The Way and won't last 6 months), you should use the browser API. Stacks completely defies this aspect of the idea, which is responsible for most of the site's accessibility bugs, and some other assorted brokenness. (The pre-Stacks site wasn't perfect, either.) In practice, this means that most frameworks should be shunned, although some frameworks are worse than others. Commented Oct 6, 2025 at 19:04
  • 6
    Part of The Way involves letting go of a lot of control over what the site looks like. This is tricky, because changes to browsers and operating systems since the late 90s have taken away a lot of user control, and I don't think we'll get it back until at least 2030 – but the groundwork is being laid, and there are ways to simulate a user having provided preferences (even if most real users can't yet express them in ways that are respected by their user agents). A properly-designed website should work in every browser, but trying to make a site work in every browser is not following The Way. Commented Oct 6, 2025 at 19:10
  • 2
    Interesting, @wizzwizz4. This is interesting, though I have to admit it's rather nebulous. I suppose that's part of the point. In a sense, what you're talking about is highly speculative in nature. I'll mull this over and maybe share it with some folks to see what they think - thanks for the helpful elaboration, I really appreciate it. Commented Oct 6, 2025 at 19:28
  • 2
    @Slate What I'm describing is very concrete and grounded, not speculative at all: it's just hard to communicate. What I've written is, indeed, nebulous. Commented Oct 6, 2025 at 19:30
  • 8
    Point-of-information: a CSS variable ("custom property") based on a REM unit remains responsive to changes in the base REM size, even when the REM unit is nested inside a calc() function. It will respond both to preferred font size and browser zoom. (Pixel-based values respond only to zoom.) SE's current theme variables are particularly unresponsive because the CSS rule (re-)defines each variable on each element, instead of letting them cascade. Commented Oct 6, 2025 at 23:09
  • 1
    Accessability is about, well, making sure the site can be used by a wide auidance - so as long as the buttons on the website look and act like buttons, and are defined with proper semantic HTML, then I would think that counts as accessable. Allowing a user to change the look of a button goes beyond accessability into the realm of "just wanting to customize things", and it's not clear that it's worth the effort to provide support for that sort of thing. Maybe some people's preference would be for all websites to look like their computer's theme, but that's just preference, not accessability Commented Oct 7, 2025 at 15:12
  • 4
    @ScottyJamison The buttons on the site do not consistently look and act like buttons, and are often not defined with proper semantic HTML. If you do that, and you define your CSS properly, then you get extras like "people can change the look of a button" falling out as emergent capabilities. (And, no, there are actually some people who need consistent UI in order to understand what's going on: most people don't need that, but it's not "just preference".) Commented Oct 7, 2025 at 15:16
  • Do you have any references to show that people struggle to recognize buttons when they don't look exactly the same? And any references about how building websites that supports swapping components (such as buttons) makes them more accessable? I would like to read more about it. Commented Oct 7, 2025 at 16:02
  • 2
    @ScottyJamison An extreme example of CSS customisation is the "reading mode" of browsers which is considered an accessibility feature. Commented Oct 7, 2025 at 17:11
  • 2
    @ScottyJamison I'd recommend the WAI's WCAG standards, and the resources linked therein, as a starting point. (Avoid the Authoring Practices Guide: it's got some correct stuff, but the overall philosophy is wrong, and that means it gives bad / counterproductive instructions in places.) Commented Oct 7, 2025 at 17:37