April 29th, 2005
I've been having fun with Google Search History this past week and have noticed a few fun things:
There is some form of search query clustering. If you look at your results page, some links will have a 'Related History' link next to them, taking you to a page of similar searches that you've done. At first I thought they were just matching basic keywords, but then I noticed some pages coming up that had no matching keywords at all (in the page title), but were still very similar in subject matter. I'm curious to see what other people's results look like. If you're curious to see what this looks like, simply click the associated thumbnail for a view of a cluster in my results.
This is the big one - previous search results embedded in-page. I notice this popping up sometimes, in two different forms. The first: If you've visited a page before, from a Google search result page, a timestamp will appear next to the link, letting you know that you've visited it before. This, in it itself, is invaluable. There's no longer a need to keep your search history, just so that you can remember which pages you've visited. Second: If you've visited a page frequently enough, it will appear at the top of your search results in a special box (much like Google News, Froogle, etc.) - this is the fundamental application of Google Search History. I want to be able to do a search and see the best result right at the top - and this application helps to make this happen. At first I didn't think this happened, but then I saw it popping up (randomly) and I'm quite pleased at this change. This is why I think GHistory is awesome.- You only have to log into one Google service for the search history to be activated - and this is good, otherwise I'd always forget. One of the first things that I do when on a foreign computer is log into Gmail, frequently before I do any searches on Google. Thank goodness for this - it really makes it a comprehensive and useful service.
Tags: google, search, history
Comment on 'Why Google Search History is Awesome'
April 20th, 2005
This tool goes through your current Google Search History, grabs all of your recent searches and turns it into an RSS feed. Would work best set up as a nightly/hourly cron job, redirecting to a file.
This tool is written in Perl and uses a few, slick, modules: WWW::Mechanize, XML::LibXML, and XML::RSS. I was influenced by the very nice webscrape tool when building this.
A sample, from my searches, can be found here:
http://ejohn.org/apps/ghistory/google.rdf
And how it looks in my newsreader (Newsgator):
Downloads
Tags: perl, google, search, rss, popular
8 Comments on 'Google Search History RSS'
March 12th, 2005
This past week, Google updated their Desktop Search software to much acclaim. The most notable update that was given (in the eyes of developers) was the addition of a programmable API to plug into this search utility. In a nutshell, it is now possible to index any file type that you desire and then search this content along with the rest of your personal information.
Giving developers this much access to a user's personal files will certainly create some interesting results. The first possible add-on, that I thought of, would be some form of a local network file search engine. Think of it this way: User A adds users B and C (other users on the LAN) to his personal network. A new main search page is constructed (by the add-on) which provides a streamlined search interface for not only user A's mulitmedia files but B's and C's too. In the search results, links are provided to the share location of B's and C's files, making them easy to access. If this were built, it would be huge. One of the major problems with doing file sharing on local network (with many users) is the ability to quickly locate and find files of interest, with Google already developing the tech to do the hard work, the rest should be easy.
Note: I haven't tested this new version of the software yet. In the old version you could only search your primary hard drive, and no others. If this has been changed so that you can search all of your hard drives then the above idea becomes super easy, without even the need for an add-on. All you would have to do is mount all of your favorite fileshares as a letter drive (X:, Y:, Z:, etc.) and then add that drive to be indexed. There you go! Your friend's files will now be indexed along with yours and it'll be dead simple to copy/use them.
Tags: corp:google, desktop, file, p2p, search, sharing
1 Comment on 'Google Desktop File Sharing'
