One of the many discussions on HATT this week revolved around Web standards. It's long been known that RoboHelp HTML (from eHelp > Macromedia > Adobe) is non-standard and non-compliant. Someone asked if this really caused a problem on end-user systems.

The answer is, of course, "it depends": It depends on the browser you're using. It depends on how you access the output (dial-up or high-speed; browser or mobile device or Internet-enabled refrigerator). It depends on the phase of the moon ;-)

My answer onlist was that non-standard output can be an inconvenience, one that you don't necessarily know about ahead of time. You could visit a page from one system today and it works fine. You could visit it from another system tomorrow and it doesn't. Or a browser could implement a change, and the non-standard site stops working.

Then someone pointed group members to Google's Web Authoring Statistics. I've read some of the earlier studies that they quote, and I was fascinated by those results. But Google's results really point out the wide variations used in Web sites today. They analyzed just over a billion documents for elements, classes, headers, metadata, and more.

(I disagree with one piece of information, however. The Google team states that "Most people (roughly 98%) include head, html, title and body elements. This is somewhat ironic, since three of the four elements are optional in HTML." Later, they state that title is the only required element. Given that title belongs inside the head tags, that implies that the head tag is also required. I learned long ago that all HTML pages required the html tag, and that the only time the body tag wasn't required was for frameset pages. Maybe they got the information backward, and the statement is supposed to be that only one of the four elements is optional? This doesn't work with their conclusion, though. More research is needed ;-). )