Not sure if anyone else noticed, but the Google homepage has also been updated for Halloween. I know, I know, I've got way too much time on my hands, but it's always fun coming across these gems.

Clickable link: http://www.google.com/ (look closely at the 'e')

EDIT: Well, that sarcasm came round full circle. There's actually a gem in clicking the logo itself. Well played, Google.

Clicking multiple times, even. I was rather amused.

Check out the HTML source: <img id="logo" width="320" height="138" border="0" onload="window.lol&&lol()" title="Click or treat!" alt="Click or treat!" src="/logos/clickortreat4.gif"/>

I used to frequent google.com, but modern web browsers let you search right from the URL bar, bypassing the site entirely. So you do have to go and look these days (or write a script that loads the HTTP and compares it to the standard version).

How do people find these things? Does somebody religiously read the Google robots.txt?

Matt Cutts posted it on Twitter: http://twitter.com/mattcutts/status/5301611018

They did something similar last year. Edit: http://blogoscoped.com/archive/2008-10-31-n76.html

And how did someone discover that?

My guess is that it was "leaked" by someone working at Google.

It wouldn't surprise me if there are SEO people watching Google's every move for hints of changes in Google search. Though I'm not sure how robots.txt would apply.

More likely than SEOs, bloggers are probably looking for any new hint of an as-of-yet publically-invisible service, so they can be the one to start the rumor.

It's too bad that /tricks and /treats don't go anywhere :-/ (I even tried altering my user agent to "Kids")

    $ wget -U "Kids" http://www.google.com/treats
    --2009-10-31 07:00:37--  http://www.google.com/treats
    Resolving www.google.com... 64.233.169.105, 64.233.169.106, 64.233.169.147, ...
    Connecting to www.google.com|64.233.169.105|:80... connected.
    HTTP request sent, awaiting response... 404 Not Found
    2009-10-31 07:00:39 ERROR 404: Not Found.

Oh well.

Very first thing I tried as well. I really expected to find something there as I would seem that any googler that thought far enough to hide such a thing in a place as arcane as robots.txt would surely finish the prank...

dupe - http://news.ycombinator.com/item?id=913137 ;)

good job adding date to make it past the darn HN URL checker.

How much of a dork/nerd am I that I actually LOL'd?

OK, how did someone notice that?