Slashdot Mirror
Preview Google's New Search Results Page
ubermiester writes "Ars Technica demonstrates how to copy and paste a bit of JavaScript to preview a facelifted Google. Ars points out that 'the changes are minimal, but they give some insight into Google's plans.'" I thought we already knew those: world domination.
Ummm, they've been experimenting with this for about two months now. I get a results page in this layout about once a week or so.
This guy's the limit!
For those who made the change and want the old way back, it's as simple as clearing the google.com PREF cookie in your browser. If you don't know how to clear a single cookie, then clear all of them and it will be included.
Google beta tests new features on a subset of its regular users, who are identified by cookies. So someone released their cookie which you set with javascript, and voila: the features show up for you.
Not everyone gets these features (so it is news to most), and your mom didn't get the new features because of a wide screen.
See my other comment for an explanation.
It appears that the green bars beside the other categories don't actually portray the number of search results for the other categories. They remain the same from search to search, even when a given search returns no results.
Just delete the google PREF cookie from your browser.
Some quick math:
I analyzed the source of the result page for "w3c recommendations". After looking at the page in the validator, I decided the following things were missing:
Two <img> tag src= attributes. Assuming a three-letter filename, that's (src="xxx.gif")*2 = 26 bytes.
Three <script> tag type= attributes. That's (type="text/javascript")*3 = 66 bytes.
Two <style> tag type= attributes. That's (type="text/css")*2 = 30 bytes.
205 attributes in the header and footer with missing quotes. That's 410 bytes.
12 attributes in each result with missing quotes. That's 10 results per page = 240 bytes.
Grand total: 772 bytes per page of results.
Using the highly non-authoritative figure of 200 million queries a day, that means that Google saves almost 144 GB of data transfer EVERY DAY, just by leaving out the quotation marks. I imagine leaving out all the unnecessary whitespace saves them at least double that.
So yeah, probably for bandwidth.
For security, the MD5 hash of this message and sig is 09f911029d74e35bd84156c5635688c0.
Clear your cookies. In Fx go to Tools - Options, I think, in Fx 1.5 Tools-Clear Private Data.
Sweet,
The person whomanages PREF=ID=fb7740f107311e46 can now look at all of the search terms entered by slash dot readers.
A back door by the Department of Justice to capture Google search queries?
comp.lang.ja...ogrammer
microsoft.pu...x.avalon
microsoft.pu...indowsxp
alt.personals.spanking
soc.sexuality.spanking
aus.comms.mobile
comp.os.linux.announce
comp.lang.functional
comp.editors
comp.lang.scheme.c
Remember Turbo10.com? And how much it's useless? Well, there is another search engine that I actually find myself useing. It's called clusty.com. There are 2 really cool features of this search engine. First, it clusters stuff. Like, let's say you are searching linux desktop. You go to google and search, you get all kinds of stuff...pretty useful I agree. However, you search clusty, and the results are broken down by subject...operating system, kernel...desktop. Very cool for research.
The second cool feature is that in the search result itself, you can preview the page. Just click the magnafying glass.
Use clusty 20% of the time, it's cool!
-- A cat is no trade for integrity!