As soon as I learned about validating HTML, I became a bit of a validation freak. That day, I re-wrote my site to validate, stopped using inline HTML, started using stylesheets, and cheered as commercial websites started moving over too.
One day I realized it didn't matter. Which, I should stress, was a significant conclusion for me to reach. Back when this stuff was brand new, modern web design was the frontier that I explored the most and had the most fun with. It's what I'm best at in my career, and where I carved out a niche. I shunned sites that didn't validate as hopelessly out of touch. I gave good SEO when you were still on the blink tag.
Which is why I sort of chuckle when people try to guilt me into validation. Son, I made that argument in February of 2001. When you can come up with a better argument than "it's just best practice" or "it makes your sites future proof", (which is bullshit), maybe I'll reconsider.
But while you're missing the forest for the trees, I'm going to make sure my site works on all major browsers, degrades gracefully on less capable browsers or devices, and has functionality that meets the objectives of the site. Even if that means it doesn't validate.
--
This topic is on my mind again after seeing all of the new Webkit functionality that's coming out. They're really on a roll, yet nothing they're releasing is in a firm spec. And this is triggering a series of responses from head scratching to full throated yells of outrage. From "Why put animation in a line of CSS when you can do the same thing with a massive and bloated Javascript library?" to "THANKS FOR STARTING UP ANOTHER BROWSER WAR, APPLE!!!11111"
I think it's time to take the religion out of web technologies. IE, Webkit, and (to a saddenly lesser extent) Firefox all have ideas for how to improve the browser experience. But time has shown that great leaps of innovation will not occur waiting for everything to be put into a nice and tidy spec. The way to push amazing new technologies forward is to say "Look, here's how we think we should address it", ship it, pitch it as the formal way to do things, and wait.
If it makes it into the formal spec, great. In 7 years developers can start using it everywhere. But that's better than waiting for 70 years, which is the approximate speed of implementing a new feature cross-browser in any other way.