Tuesday, August 12, 2008

URLs Everywhere

Google analyzes backlinks posted on web pages to calculate PageRank, which in turn influences the search results ranking of pages. But web pages aren’t the only places to find URL pointers. Whether or not Google evaluates any of these, I don’t know:

• There are URLs on billboards (some of which Google already stores in Google Street View)
• There are URLs scribbled on walls using black marker
• There are URLs mentioned in books (some of which Google already scanned as part of Google Book Search)
• Besides books, there’s URLs in other print media like magazines, newspapers, comic books
• URLs are sent around in emails (some of which Google already stores in Gmail)
• URLs are sent around in chat programs (and when you’re using Google Talk, Google could analyze this)
• URLs are sometimes pasted in the chat rooms of IRC (or in the 3D worlds of Second Life, Lively and others)
• URLs are sometimes appearing in text documents, spreadsheets, presentations (Google knows those stored in Google Docs)
• Sometimes, URLs are mentioned on web pages but not linked, like in certain news reports; also, URLs may appear in Flash files, JavaScript files and so on, some of which Google say they already crawl
• URLs appear on TV in spoken or printed form; they are also named on radio
• URLs are sometimes sent as SMS
• URLs are sometimes talked about
• URLs are sometimes thought about

Whether accessing any of these makes sense is another issue. Billboards would be ads, so they would probably be excluded. Links sent in emails are often spam (though Google has ways to find out about certain spam and could exlude those mails, or conversely, add penalty pointers to the linked pages); links sent in emails are also sometimes private; so are some of the URLs talked about, or thought about (and the latter is hard to scan with today’s technology in any case). In other instances, the sample would be skewed as it would only be the “Google property” sample (e.g. when analyzing URLs in Google Spreadsheets but not in MS Excel files). Links in books could be more authoritative, though they might also be outdated. But sometimes, analyzing a URL outside of a web page may also give a better, because more diverse and fine-tuned, ranking.

URLs Everywhere by Philipp Lenssen | Comments (9)

>> More posts

Advertisement