"Google doesn't care, and these other sites don't care, so why should I care?"
I have found that validation serves a very important purpose - to guide your development. I have lost count of the times I have encountered CSS with all sorts of extra junk because things didn't look right, and the reason for it was that the underlying HTML was invalid.
Yes, the browsers are smart. But they are still guessing. If you leave out a tag, the browser will close it for you. Not always in the right spot. Invalid pages could also result in a DOM tree that's not quite right, so JavaScript could be affected.
Developers should validate their pages just as they lint their code, to ensure quality.
Of course, if you're using cutting-edge elements, they won't validate. But that's a little different story. It's not a reason to not validate, it's just a reason to filter out those errors.
Some of us still have to support older IE versions which needs to some workarounds to work with it. Even the validator.w3.org doesn't pass it's own validator. https://validator.w3.org/nu/?doc=https%3A%2F%2Fvalidator.w3....
My point is that we need to use modern tools like html5 while still being backwards compatible with older browser which means breaking some w3c rules.
Another valid case off the top of my head for breaking w3c(haven't tested if it passes validation) would be `<meta>` tags inside body tag. Which is part of the microdata format. Normally you're not supposed to put meta tags outside head but for while using mircrodata you sometimes have to rely on meta tags because your design doesn't allow for required(by microdata) to be shown on the website, so you use meta tag, which doesn't do any harm since it's not rendered for the user.
HTML hasn't changed in years, and XHTML is largely backwards compatible.
Had to point this out: windows are real, the broken windows effect is something from "real life".
Browser tech has been moving faster than validators can keep up. Newly specced features are being polyfilled, etc.
Browsers and search crawlers deal with this; probably time for validators to be about something other than the site's static assets.
This article highlights the worst of the "webmaster" mentality – ignoring the end results in favor of things like validation errors and tools used.
They trust Hillary as a politician because the source code has secret comments? Seriously? And then they trash Chris Christie for the same exact thing. And Bernie gets two thumbs down (and a shot at him for being a socialist... along with an ad for the authors) because he used WordPress? They call his site a mess but don't describe why. Bernie's site is impeccably designed and works well.
I'm sad about this article. Design and the web play a HUGE role in presidential campaigns. I happen to believe that you can tell a lot about a candidate from their website. But this was a shallow, pedantic and opportunistic waste of a potentially great topic.