Huh? There's little to search with when it comes to images, unless you pull from the surrounding content, which Google does. There are several orders of magnitude difference between storing web page-size content and the kind of filename-size content that would appear in image titles and descriptions. In other words, an index of just JPG image titles would be far smaller than an index of the same number of web pages (though the binary files themselves, which won't be indexed, would likely consume more space). -Declan On Tue, Oct 09, 2001 at 06:50:40PM -0700, David Honig wrote:
Very nice analysis.
If you greatly reduce Google's speed of search, what kind of compression-gains can you get? Imagine an archive which is highly compressed [1] but used mostly to counter censorship.
[1] That JPGs etc. are already highly compressed means that if you keep pictures, you won't gain as much by trading off search speed for compression.