Imagine a web of nothing but infinite VMS documentation!

Also:

I could easily write a robot which would roam around the Web (perhaps stochastically?), and verify the html, using sgmls. Then, whenever I come across something that’s non-compliant, I could automatically send mail to wwwmaster@sitename. No one would have to annoy anyone else about whether or not they’ve verified their HTML; a program would annoy them automatically.

Not the onion.

From such humble beginnings, this big pile of crap: http://tgs.freshpatents.com/Web-Crawler-bx1.php