Please log in.
Before you can vote, you need to register. Please log in or create an account.
Computer: Searching: Link Quality
neogoogle I   (+1, -5)  [vote for, against]
scroll bar data complements pagerank to show trusted sites that readers dwell on; noticing pages that have [control] [c] key events winnows pages to those with super high utility

pagerank is patented it works with a measure of links between trusted sites; a new idea I call Well Read complements or partially replaces pagerank at a competing site; visualize a transparent gif column right near the scroll bar; mouseover event at the transparent gif goes with scroll bar usage;

if the person uses the scroll bar once or twice that means the html page has a high readership application rank: it is Well Read; googles ad area is a module that could create the Well Read transparent gif sensor as well as pass the data back to google; also you could just use whats known as a web beacon; further the sensors could be as a grid on the page perhaps with CSS or transparent letters that have mouseover scripting

with pagerank you get what UC berkeley links to; with Well Read you get what people like enough to read fully; combined you get those pages that UC berkeley links that people actually purposefully dwell on; the most useful part of the web

here's a critics FAQ; then there's a word about microsoft

//scrollwheels a --- transparent gif grid is able to detect any vertical scrollwheel activity

//Dragging the scroll bar, clicking the scroll bar track or pointers, or using the cursor keys or page down or the scrollwheel has no effect on the image if the mouse pointer is off the window

that is fine as rather than ask every visitor to the mall what they think of build-a-bear you could just ask a few hundred to get a picture of visit length

//Mouseover gifs would also trigger mouse movement events unrelated to scrolling up or down, like moving the mouse to the 'close window' button, or back button etc - mouse movement does not have a direct correlation to attention to content

thats why the gif grid variables have different names; we care a bunch more about the right side gifs as well as the derived oscillate on page scrollwheel data

a word about microsoft: microsoft has an urge to be a search engine presence neogoogle I may give them an opportunity; there is a different opportunity which is compelling: javascript can say when the [control] then "c" key events of Copy [^c] events happen on a page; accumulating data on what pages are wonderful such that users actually copy data from them creates a very high utility ranking

if it is trust legal using the windows OS to be aware of which parts of microsoft office are up then you get data on pages that people actually copy online data from onto spreadsheets, powerpoint, or word; this creates the opportunity to make a MSN productivity plus search engine that might be measurably effective at building productivity at work

pagerank: trusted data

well read: pages that users dwell on to read

App useful: pages with data so marvelous it gets copied

those three together create very high utility search
-- beanangel, Mar 07 2008

Neogoogle II n III at my notes blog http://www.imminst....ule=blog&blogid=22&
find neogoogle [beanangel, Mar 07 2008]

mouseover works with scrollwheel http://www.dynamicd...11/gradualfader.htm
[beanangel, Mar 07 2008]

noticing when people ^c: key events that javascript is aware of http://unixpapa.com/js/key.html
the [control] key then the [c] key are detectable with javscript [beanangel, Mar 07 2008]

So a page that fits on one screen gets a low ranking? A visitor with a high-resolution monitor counts less than one with a low-res monitor? A visitor who scrolls down a page using the mouse wheel (or page-down) counts less than one who uses the browser's scroll bar? Information at the bottom of the page counts more than information at the top?
-- angel, Mar 07 2008


all that, and less: "throw exceptions"

one thing about page stats is that with this technology as a grid you can tell if the page load duration goes with actual reading as well as scroll bar activity; thus a page with a neogoogleside measure of datasize fits onscreen with a plausible view time like more than three seconds but less than three hours has an attentional rating

it is different thing than people with rare habits like [pgdn] not mattering; I just verified that the mousewheel energizes javascript mouseover events (with delay) at [link] this would go well with a transparent letter grid as well

a --- transparent gif grid is able to detect any vertical scrollwheel activity

//Dragging the scroll bar, clicking the scroll bar track or pointers, or using the cursor keys or page down or the scrollwheel has no effect on the image if the mouse pointer is off the window

that is fine as rather than ask ever visitor to the mall what they think of build-a-bear you could just ask a few hundred to get a picture of visit length

I frequently use full screen windows; this motion []<- is rarer than [->] ; the quality of the data does have something to do with the users mouse being on the page; but as this is hyperlink filled the [] area gets either a [null mouse presence] value or[mouse on page]; null mouse presence would be an "exception"; you get to toss some []<- data to accumulate a valid dwell picture with multiple users
-- beanangel, Mar 07 2008


//all that, and less: "throw exceptions//

I have absolutely no idea what this means. Also, does your mention of page load duration imply that a slow connection would give a higher ranking than a fast one? (Or is it the other way round?) The effect shown at your link is irrelevant; it relies on the image passing "under" the mouse pointer because it's a mouseover event. You say that "mouseover works with scrollwheel", but it only does so if the cursor passes over the container that the mouseover event lives in. Dragging the scroll bar, clicking the scroll bar track or pointers, or using the cursor keys or page down or the scrollwheel has no effect on the image if the mouse pointer is off the window.
-- angel, Mar 07 2008


Yeah, it's a bodge attempt at solving a real problem - how can you tell whether your content is being read or not?

Interactivity in the form of annotations and user-provided content is one method (or at least it would be, just as long as you don't attract the usual blog-spam that occasionally crops up) - although this only tells you what the annotators are doing, there may be many others who read, but don't participate.

Mouseover gifs would also trigger mouse movement events unrelated to scrolling up or down, like moving the mouse to the 'close window' button, or back button etc - mouse movement does not have a direct correlation to attention to content. Yes, it suggests your website visitors are still alive - but that's about it.
-- zen_tom, Mar 07 2008


And you won't capture scolling by wheel.

OTOH, my server knows how long my pages are viewed. I'll gladly sell you that information on a subscription basis...
-- phoenix, Mar 07 2008


[phoenix], I bet it doesn't.
-- angel, Mar 07 2008


Yeah, timeouts aren't counted, but I know when a page is shown, and when a user navigates to another page. It's safe (usually) to presume the intermediate time is page viewing time. I expect one could get around the timeout issue with some AJAX.

We don't (but could though cookies) peg page view time to a particular visitor.
-- phoenix, Mar 07 2008



random, halfbakery