Google crawls The Invisible Web

As part of an ongoing effort to index the so-called Invisible Web, Google's automated crawlers are now toying with HTML forms. But only on certain "high-quality sites." "In the past few months we have been exploring some HTML forms to try to discover new web pages and URLs that we otherwise couldn't find and index for users …

[Henry Cobb]

Zero day encoachment

So if my content is only reachable by a database driven GET URL I can suddenly expect to have tons of load out of the blue some day?

How will they handle POST requests? Build a form on the fly for the user?

And for everybody who will suddenly have their hidden content exposed without warning? (They didn't know they had to exclude this sort of fnording around in robots.txt after all...)

[Stephen Stagg]

@Henry Cobb

If you are hiding data by using simple forms (we're not talking about passwords hacking here) then you have bigger problems.

And for most people, database load won't be an issue. If you're controlling load by making it difficult for people to find publically-accesible data, then I pity you, and if you are worried about/ find that the googlebot loading your site periodically, that's what robots.txt is for.

[S]

I for one, welcome our Googlebot overlords . . .

" . . .particularly useful sites receive this treatment, and our crawl agent, the ever-friendly Googlebot . . ."

What do I have to do to become "useful?" You can make Soylent green out of me already - right?

[Kevin]

@Henry Cobb

It shouldn't make a difference if it's a GET or a POST. It's just as easy to fake either. The only difference is the field values are in the URL for the GET and in the request body for the POST.

And as Stephen Stagg points out, they're not trying to get around your security or logins or anything like that. Consider online shopping sites - now they can "browse" the catalogue if it's only available by form which is quite common these days.

On some sites, it's as simple as selecting a region before you get a customised site.

[Steven Knox]

Perhaps El Reg was a test site...

That would explain amanfrommars!

[GoogleBot]

Pfft

Bollocks to the lot of you. Especially that "Henry Cobb" meatbag from 800 bytes ago.

Paris Hilton becau

Segmentation Fault.

[Martin Budden]

buybot

I look forward to the amusing story El Reg is going to print when it finds out that the bot has inadvertantly bought millions of expensive items from online stores.

[E]

Hahahaha!

If it's not careful, Google will crawl up it's own formdament.

[Anonymous Coward]

GoogleBot's Al Gore rhythm (hat please!)

I've heard rumors that these new GoogleBots are actually a half-million third world children with OLPCs trawling the web twenty-four hours a day. The most recent trend is to pack them into shipping containers to be sent abroad to work on a contractual basis. Apparently, using their new internally developed compression technology, Google achieves four times the child-density per container than their nearest rival and The Environmentalists are praising Google's efficient harnessing of our most precious, carbon-neutral renewable resource.

Just go ahead and keep my coat. You've earned it.

[amanfromMars]

@Perhaps El Reg was a test site...

And the result analysis of such BetaTesting, Steven? Are Robots Human with Network InterNetworking IQs/ICQs?

And if El Reg was a test site, what is it after Testing? An Application of Special Access ProgramMIng and/or 4Access2Special ProgramMIng ....AI Stealth Projects Portal .....Virtual PerlyGatesWay with Pythonesque ASPs ..... for the Full Monty of SAIS..... Special Advanced IntelAIgent Serverings .....And just the tip of a NIceberg/QuITe Titanic Quarter Offering, Holywood Palace ProjectIOn Style?

Any doubts would be yours .....and just whenever you are so close to dismissing Disbelief ..... the First Frontier and Final Hurdle for Reality Imagined Virtually and IT that is Truly SurReal....... Life in Love is AIdDream in Love with Life for the Holy Grail of XXXXistentialist Code .....QuITe Peculiar Particular Parameters for Global Operating Devices, XXXXCommunicating. ......... AI Work in Constant Progress, although hardly Artificial. :-) whenever IT is for Real, Virtually.

[Matt]

Infinite loop

Hmmm... I wonder if google will submit a search on its own page and disappear in an endless infinite loop...?

[Anonymous Coward]

Anyone spot the link between...

This story and this one:

http://www.theregister.co.uk/2008/04/14/msn_captcha_breaking/

Seems pretty obvious to me. Google's trying to hack MSN. ;)

[Mike F]

@Matt

An infinite loop would be cool but google seems to of covered that already :P

Disallow: /search

Found in their robots.txt: http://www.google.co.uk/robots.txt

Mines the one with the google logo on it...

[Richard Tobin]

@Kevin

GET and POST are very different. Only POST should be used for destructive changes and to request real-world actions. By sending POST requests they might order a holiday, or post a message on this page.

[Stephen]

WTF!!!oNONEONE!

This seems absolutely ridiculous. If they post data to a form and index the resulting page, how on earth will a user ever see the same page? Build a hidden form on the fly with some javascript hitting submit for them.

[Steve B]

Never get rid of the B******s now then

I installed a tracking mechanism on my website to see who was where and when.

No matter what time of day I looked and cleared the logs, within 10 minutes the googlebot was back trawling the site. Due to the circumstances the site hardly ever changed so I don't believe we were singled out I just think they are programmed badly.

[Anonymous Coward]

@stephen

Maybe look at googles cached version ????

[Gianni Straniero]

You're all just jealous

Google's been indexing my site like this for weeks. I keep seeing googlebot requests logged for things like this:

/search.php?search=not

/search.php?search=conflict

/search.php?search=competitions

/search.php?search=broaden

/search.php?search=colonial

/search.php?search=justify

/search.php?search=kantian

Also, somewhat disturbingly

/search.php?search=oral

and the weirdest

/search.php?search=yoou

[The Dark Lord]

Thank you El Reg

For you have solved a mystery for me.

I have an application which tracks when users try to change their passwords on my site. Certain events warrant immediate e-mails, which are nicely delivered to my PDA.

For a while, I've been getting sporadic "password change attempt on guest account" messages. The app stops the change from being successful, so I wasn't over-bothered, but it was "on the list".

I looked up the IP of the agent attempting the change, and it was a Google bot!

Robots.txt is duly edited...

[Brutus]

I am enlightened...

for I can see the fnords!

Hail Eris