If Google can read .css does this mean robot.txt (with disallow) will be useless? Im confuse please enlighten me
Using Mozzila browser you can already have thumbnails next to your search results in google, so this is not new either. I have thumbnails in my search results and have done so for a couple of months, the reason i do, is so that i can view what the site might look like before i click the link, which helps to spot some of those spamy sites straight off.
You do as an extension of your browser, yet not part of the WEB PAGE you're viewing (like this person speculated Google wants to). The difference is significant, because the search engine results pages views with thumbnails would be consistent across browsers, regardless of whether or not someone had that extension installed. No, it's not new...other pages have done it.. Alexa does it, etc...but it would be new for Google to do.
I don't understand how detecting hidden text requires accessing stylesheets. I mean, I know that stylesheets can be used because of markup to help hide text, but trying to detect something, there's no gaurauntee of a stylesheet being accessed. That's kind of like when ya go through the airport...they try to detect bombs, but that doesn't mean they look in your pockets... even if you empty them, that doesn't mean that they've looked in them.
There are several problems I see with this (besides any proof of them doing it). Is this used to somehow find spammers? (i.e. hidden text, etc.) Heck, I can give them a list of straight forward techniques that they can't even eliminate yet. Is it used to present better AdSense Ads based on where the elements is displayed on the page? Is it used to provide better rankings by somehow, for example, figuring out where the page elements are and giving less "value" to navigation or footer links as opposed to links within content? The question of intent is important because it also somewhat determines the level of sophistication they would need. Simply capturing a screenshot for display purposes in search results is no big deal. But accurately analyzing screenshots is a whole other ballgame. Think of the little security word verifications used when you login to blogger or many other sites. Now think of trying to dynamically analyze an entire page as a "graphic" not text, especially when shades and colors could be very close. Let's say they are trying to find hidden text. Isn't some hidden text valid? How about a webmaster who has some code in his markup that he is working on but it hasn't gone "live" yet? Is he spamming? How about a mouseover that would tigger display text or a description in a table cell or div tag? Now you are talking about not just reading and rendering CSS code or analyzing a graphical snapshot, but also executing javascript and functionalities of a website as a user would. And would their bot be compliant with all forms of javascript that work in different browers? JS is even more browser and version specific than CSS is!! Even assuming it is all possible, is it economical? Are you saying this is going to be done on every page on the ENTIRE internet? If this is true, I hope the devil has his iceskates ready.
Googlebot 2.1 is hitting a site with both external css and .js ... but the bot is not downloading them: HTTP/1.1 Mozilla/5.0+(compatible;+Googlebot/2.1;++http://www.google.com/bot.html)
Hitting a page that HAS CSS and Javascript is what they have already been doing for years. The question is if they are downloading, executing, and analysing the results of javascript and CSS as it renders in a browser. I have yet to see any proof of this but I would be very interested if anyone has any.