Hacker News Reader

I agree. Google simply followed the the semantics of http. They had similar problems with the the Google Web Accelerator before. But it's really hard to blame google for the mistake of other programmers.

Additionally, the interface to web crawlers (robots.txt) is well defined.

Yes, it is an error. What alarms me is Google's attitude.

What happens when we're dealing with interfaces that are a little bit more ill-defined? Will Google continue to demand that you follow their way of doing things?

Google's attitude in this case suggests they will.

It's not "their way of doing things", it is just the way the web works (per HTTP spec). Any crawler would have done the same thing in that situation -- that fact that it happened to be Google is merely coincidental. Given the scale at which they operate, you can't expect Google or any other web-scale crawler to be mind-readers.