Has anybody ever taken a look at the html of google? A few months ago i took a look, and i was shocked that a website as big as google could have html as bad as that which was shown.

The index page alone holds 49 errors when validated according to html 4.01 transitional. html 4.01 transitional is used by default when no doctype is specified.

Do a search for "forums" and validate the results page, and it has 280 errors. Again, no doctype is specified.

Run the validator against search.msn.com and it validates perfectly. run it against a results page for the same query: "forums" it validates perfectly.

http://validator.w3.org/check?uri=http://google.com
http://validator.w3.org/check?uri=http://google.com/search?q=forums
http://validator.w3.org/check?uri=http://search.msn.com
http://validator.w3.org/check?uri=http://search.msn.com/results.aspx?q=forums

Why couldn't a company as huge as google code valid html?

Well, it works for them doesn't it. I've never seen it display wrongly in any browser, ever. Who cares if it validates?

Why couldn't a company as huge as google code valid html?

Better question is - why should they?

Designers seem to get really twisted in knots whenever they see a big site that does not use valid HTML - so here's a couple more for you:

www.yahoo.com
www.msn.com

It doesn't matter

http://validator.w3.org/check?uri=http://www.webhostingtalk.com

Errors: 1607

Yea, Just because a website doesn't follow strict HTML markings as set by w3, doesn't necessarily it's bad coding; It's more of an "image" if a site is HTML/XHTML/CSS valid which doesn't necessarily mean it won't work cross-browser.

Google has a simple look and works evry where in each browser!
Why they should make them site W3C validated...

My opinion: It´s a shame, but I understand!

But I hate it when I see a site who sales hosting or webdesigns isn't W3C-validated! Argh...

Google may very well have many of its own reasons for not creating "validated" code. Keep in mind that Google runs proprietary operating systems, software, and hardware configurations. They have to serve up millions of requests and must optimize their code accordingly. I read a more technical article about the optimization stuff somewhere, don't remember or want to go into details.

So as long as their pages are cross-browser compatible and loads quickly it doesn't really matter if it's validated.

ph23man is right. Another thing is, people can view Google easily even if they use a text-only or outdated browser. A lot of people forget this about standards, that some sites need to be viewed by people on older PCs, not just the tech-savvy crowd.

There's more to it than all this even, and Mikey brings up some good points in his critical observation.

A few of us got together and created a cost-benefit analysis of the bandwidth google expends with its current site and how much they could potentially save if they reengineered their site for compliance and best ADA practices. We actually retemplated the main areas of Google, rewrote some of the JavaScript they use from time to time, and at the same time did some math on their current usage and costs.

We are preparing a case study on this to present to them that shows a seven figure savings per year, based solely on client-side code reengineering. This is not a novel concept. The same has been done with msn.com, and after someone did this with ESPN.com, they went through the reengineering process. Staples.com is going through the same process.

Validation only really matters inasmuch as the site needs to work for the widest audience possible. Unfortunately, poor engineering on Google's part makes the site unusable for my good friend Herman, a blind man who recently discovered email and the Internet when he made the mistake of letting me into his apartment one day :) Google is utterly a waste of time because its poor engineering makes it so it takes 5-10 minutes to read one page of search results, and he can't even skip between them using keyboard functions on which he's come to rely.

So, there's a rather large market for a global company completely wasted. And that's just one example. The ROI analysis we did doesn't even cover possible returns if Google makes its site more accessible to people with disabilities.

As for the statement about not using W3C markup not meaning code is bad, actually, that's exactly what it means. Browser manufacturers have all committed to W3C as their standard for production. W3C recommendations have been adopted by every browser in the industry (yes, even IE), and not adhering to W3C is, by its very definition, bad code, according to industry standards. Sure, bad code can still be functional, but it becomes the responsibility of the browser to try to interpret what you meant when you used it. You don't really want to give that responsibility to the browser, because what works today might not work tomorrow. Now, you can write valid code that's absolutely dreadful. I see it all the time. Valid does not necessarily mean good. All it means is that you're coding within the industry guidelines. But it's a baseline requirement. It's a starting point.

Originally posted by the_pm

A few of us got together and created a cost-benefit analysis of the bandwidth google expends with its current site and how much they could potentially save if they reengineered their site for compliance and best ADA practices.

That would certainly be interesting - especially as addition of all those commas for individual values, ie, border="0", could add a couple of k to the page - and with Google server metrics (http://www.platinax.co.uk/news/archives/2004/12/google_server_m.html) as high as they are, that couple of k could add very substantially to overall bandwidth usage and costs. Neither of which W3C takes into consideration.

As pointed out before, just because a site isn't W3C compliant, doesn't mean to say it doesn't have excellent cross-browser compatability. And just because a site is strict or transistional doesn't actually mean to say the site is useful or good.

W3C compliance is a good goal to aim for, but there are always wider concerns with a site than simply coding compliance issues.