A reminder that, as article points out, a healthy web needs multiple independent client implementations and we're already down to less than a handful major ones. Ideally they should hold equivalent usage share which implies that as currently usage is heavily tilted towards Google Chrome, the best way in which you can help Mozilla is to use their browser and get your friends and family to use it too, unless there's a…
...and ideally, we should also encourage the use of the "lesser" browsers like Dillo, NetSurf, and all the text-based ones. They can't really run "web apps" and the like, but will be fine for viewing the "long tail" of content-focused sites out there (including this one.) What has become quite obvious to me within the past few years is that the whole "move the Web forward" thing seems to be really about coming up wit…
Yeah, it's enough to be infuriating. I've recently been thinking about what it would take to implement a lynx-like browser that had a better understanding of layout and could deal with the modern web.
Hurdle number one is that there's no "Web standard", but rather a great many interrelated standards. At a bare minimum, you need HTTP K₁, SSL/TLS K₂, HTML K₃, XHTML, DOM N, CSS K₄, and ECMAScript K₅. Several of those standards are actually multiple standards, so you'll have to do some digging. That's the easiest hurdle to overcome. Number two is that a few of the standards (including some really fundamental ones like the DOM) are maintained as "living documents", making them moving targets; even the—uh—"dead" documents are subject to speedy revisions (for example, there have been 3 new ECMAScript revisions in the last 3 years). Number three is all the up-and-coming experimental stuff that isn't standardized yet, but if at least one browser supports something, some website somewhere is already using it. Between two and three, you'll need to keep a finger on the rapid pulse of the Web. And on top of all that, you have to support legacy bullshit, deal with malformed documents, and be ever hyper-aware of security threats.
Allowing for off-the-shelf components like libcurl and duktape, I'm sure it's still doable by a single person, but it's a lot of work (much of it frustrating).