Google Desktop
I’ve been using Google Desktop for two weeks now, and its nice. Like most of the other Google products, its simple to a fault. Yes, there’s almost nothing to configure and the interface is a single text field entry. A no-brainer. It indexes all local hard disks and email automatically, and there’s little options of modifying its behaviour.
Could it work without the famous PageRank algorithm that powered it’s web search? Yes, most certainly. It has some algorithms that rank the relevance of results. For example, it tracks when each file/email was accessed, and it ranks most recently accessed results higher. There are probably some other indicators it uses for ranking, but they are not obvious at this time.
Compared with Lookout, it tracks local content better. It searches web history and all files, and does it in real time. There is a small process that runs in the background that indexes content on-the-fly. It lacks Lookout’s mail filtering rules - such as limiting searches to certain folders or within a time range. It however does process “To” and “From” fields.
Unlike the web search, it’s “hit rate” is far lower. I often refine my searches and browse through the results to find the email/file I was looking for.
It’s content ranking could do with more parsing of the content, such as identifying the headers or file names. I admit it’s hard to read my mind and understanding and organizing my (mostly) unstructured data will be a challenge worthy of Google.
I’m quite happy to live with another addition to my desktop - it’s small, fast and works reasonably well.