1. 32

  2. 17

    Long story short, if you care about the future of the web even a little bit stop using Chrome and switch to Firefox. The only way to keep Google in check is by having an independent web browser that has enough market share that people care about it. If Google manages to kill Mozilla we’re in for a really bad time because an ads company is going to become the gatekeeper of the fucking internet.

    1. 2

      At this point, I am wondering if Mozilla is part of the problem. They were instrumental in bringing down W3C which was the standards body concerned with the HTML standard. If W3C was still relevant, this would not have happened.

      1. 1

        Mozilla didn’t bring down the W3C, the W3C did. After HTML 4.0, they stopped listening to the needs and desires of web authors and browser vendors, and started working on their own blue-sky projects that nobody actually needed or wanted, like XHTML and XForms. If W3C still controlled the definition of HTML, either it would be basically the same as WHATWG in our timeline, or else it would be irrelevant in a world of App Stores.

        1. 1

          I do not know a lot about the needs of other HTML web authors than my own, and hence can not comment more on that. However, I have worked with URLs, and I do not see how the additions by WHATWG are actually useful when one compares it to what existed before. The reason for the existence of a standards body is also to say NO when needed, and to oppose half baked features.

          1. 1

            Sometimes standards bodies work prescriptively, where everybody comes together to agree how a thing should work before it gets implemented. In such cases, opposing half-baked features is vital and important work, even though it’s not always successful.

            Sometimes standards bodies work descriptively, where a bunch of contradictory implementations already exist. In such cases, the standards body has to pick the most widely-implemented behaviour and document it, even if that implementation seems stupid. Occasionally a standards body will say “we can’t standardise something as ugly as this!” and come up with a nicer approach; in that case, implementors will try implementing things the nice way, howl in frustration when it doesn’t interoperate, and wind up implementing the ugly behaviour anyway.

            I’m not sure what particular additions to URLs you’re referring to, but I recall hearing that the LDAP standard had a “URL” data type that exactly matched the original RFC, and therefore was a subset of what WHATWG called a “URL”. The LDAP community perceived WHATWG as seizing control of “URLs” and trying to redefine them, but WHATWG didn’t create the mess, they just found it and were trying to understand it.

    2. 13

      It doesn’t really matter if we had community processes or not…baroque standards and byzantine specs create evolutionary pressures that cannot be ignored.

      If a slow-but-basically-compliant web browser can’t be hacked together by an intern in a summer, and if the scope of compliance is ever-increasing, on a long-enough (but shorter than we think) timescale only large companies will ever be able to build browsers.

      So, um, maybe start actively fighting to cut down web standards to manageable sizes.

      1. 2

        So, um, maybe start actively fighting to cut down web standards to manageable sizes.

        Do you think that’s possible? I find it hard to believe that we can turn back the clock here. I’d love to but it just doesn’t seem plausible. I guess that means we have to start again somehow, which also seems implausible.

        1. 2

          I think it might be possible, but I also think that starting again–at least for the subset of users that care about such things–is probably easier.

          A good chunk of the web complexity is layers of cruft built to support one another and to explore a design space that we are all pretty familiar with now. I imagine that a Pareto browser wouldn’t be as gnarly as we think.

          1. 0

            I’m enjoying this thought experiment :)

            Where do we start again from? An older/simpler HTML over HTTP 1.1? Something that mothra can render? Or do we build from gopher?

          2. 2

            Regarding “a slow-but-basically-compliant web browser”, I’d love to see how much browser functionality can be “polyfilled” by libraries that are generic enough to run across various browsers, cutting down how much needs to be implemented by a browser itself.

            For example, if a browser implements JS DOM manipulation, would it need to bother implementing CSS, or could it just inject a generic JS script for handling CSS? What if we only implement WebAssembly, and use that to run a JS interpreter? I imagine implementing canvas would allow lots of rendering-related things to be offloaded (images, fonts, layout, etc.) although personally I wouldn’t want to lose the text-first nature of the Web (e.g. selection, copy/paste, search, etc.).

            1. 2

              This is a cool idea! I can imagine a cut-down web browser that is basically just a sandboxed interpreter with a HTTP client and a drawing area (OK, and local storage, and audio, and…). A browser engine controlled entirely by scripts, like luakit, but… more.

        2. 12

          Jen Simmons put it well:

          Think of all the many HTML elements that were considered and rejected over the years — and we are supposed to be on-board with TOAST? Because a couple guys at Google decided they want it. And they can. So no to <footnote> <author> <publication-date>

          But yes to <toast> ???

          1. 10

            I try to be measured in my responses, but <toast>? Seriously? Because it pops up? I get that it’s a fairly domain-specific concept so it’s gained its own term (I guess?) but I can’t think of another HTML tag named using such a bad, unintuitive analogy. Why can’t cutesy stuff like this just go in a web components library?

            1. 18

              I … never made the link between “toast” and “popping up” until your comment. I always thought it was a weird name copied from Apple or something.

              As an aside, I feel many “toasts” are not good UX: stuff like “disappear after 3 seconds” can make it hard to read and impossible to copy, especially for those that have poorer eyesight, are less literate than the average programmer, not native English speakers, have motor skill problems, etc. It also often obscures important UX elements, which can be annoying. I’m wary of codifying something like this in the web standards.

            2. 3

              I’m a bit lost, what is <toast>? This? https://github.com/jackbsteinberg/std-toast

              Oh here’s something: https://developer.android.com/guide/topics/ui/notifiers/toasts.

              (Yay android docs don’t scroll if they can’t load javascript from gstatic.com.)

                1. 3

                  I was surprised to run across a good comment on that article, about fire-and-motion in regard to chrome.