add, search, annotate, link, view, overview, recent, by name, best, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Semantic content analysis read-ahead from a webpage determines the overall intelligence (or more likely stupidity) of each link and changes the link text to a brighter or dimmer colour, depending.
Reminds me of...
Net_20Cultural_20Ev...0Indicator_20Theory
What you get when you track the summary of these... [RayfordSteele, Sep 16 2009]
[link]
|
| |
Similar in concept to bing.com. More immediate and helpful, it seems. How would the engine evaluate intelligence? Would certain words such as "Youtube poop", "lolcat", and "Guitar Hero cheat codes" trigger a negative result, where more "intelligent" words yeild a more positive result? |
|
| |
I just noticed this idea, for some reason the link was barely visible. |
|
| |
It would be the ultimate spam filter if you could perfect this technique. I think it's overly optimistic given current capabilities. |
|
| |
spelling, ratio of upper to lowercase, number of exclamation points for content; for a search-engine flaggable things like the search phrase actually being part of the visible text, lack of recursivity in links, document size... stuff google shoulda put in but didn't. |
|
| |
//number of exclamation points for content//
Yes, very good. Bun. |
|
| |
Nice concept. I'll take the Grammarian package. |
|
| |
I agree that it's overly optimistic. To preread every link to such a
critical extent would mean waiting a long time for your search
results to populate. Then again... it might be worth the wait.
Maybe have a numbering system for the intelligence level, and
be able to filter your search for a specified range? |
|
| |