Google Brings Search to the Desktop

Information Management Journal, Jan/Feb 2005 by Swartz, Nikki

Everyone who uses a computer today must manage, organize, store, and then try to recall where they have stored their documents, photos, music, e-mails, and other files.

Google hopes its latest invention - Google Desktop Search - just might make the job a little easier.

According to Google, bringing search capability to the desktop was the number-one request from its users. Desktop Search can search any Web page a user has ever visited, any e-mail they have ever opened, and the transcript of any instant message they have ever sent or received. It also can find files by searching for words inside files, which is helpful when PC users cannot remember what they named a file but can remember what it was about. Google Desktop essentially becomes a sort of "black box" for a computer, giving the machines a "photographic memory," Google says. When users conduct a Google search, the query is sent simultaneously to Google (to search online) and to Google Desktop (to search the PC). Users will receive search results that include Internet sites, e-mail messages, and document files.

The program sounds good, but there are some technical and privacy hurdles. Officially in beta testing, Desktop Search can't search Acrobat (PDF) files except by file name. It can't search Web pages visited unless Internet Explorer is the browser being used; chat sessions unless AOL Instant Messenger is used; or e-mail unless Outlook or Outlook Express is used.

The entire program fits in a 446-kilobyte download, but installing it requires at least 1 gigabyte of free hard-drive space because it works by creating a multi-megabyte database or index of the words in all of a computer's files. To search such vast amounts of material, it needs a healthy amount of space for its index. So Desktop Search requires between five hours and one entire day to build its index. Once it is built, Google maintains it by logging every document and message opened, every Web page visited, and every instant message session conducted.

Google searches the computer's main hard drive, but the company says it does not see desktop-only queries and their results, and because Desktop Search does not require an Internet connection, Google cannot access the giant index maintained on the user's computer even if it gets a subpoena to do so. Also, users can turn off any of the searchable item functions and omit secure Web pages from the log, so that their banking and stock transactions cannot be recalled, for example.

Still, Google warns managers of shared computers to think twice about installing the software in computers at libraries and Internet cafes until it develops advanced features such as password protection and multi-user support. Users could unwittingly allow people who follow them in using community PCs, for example, to see sensitive information in e-mails they've exchanged. For example, typing in www.hotmail.com on a Desktop Search-enabled computer will reveal copies, or stored caches, of messages that previous users have opened. Users can exit the program altogether, preventing it from recording Web surfing, e-mail, and chat sessions. Still, even if a user deletes an e-mail or file, a copy remains on the PC's Desktop Search index, and this information could be discoverable in litigation.

Despite these challenges, some Web analysts have called Google's innovation the most significant new Internet technology in years. Microsoft, America Online, and Ask Jeeves say they will release similar desktop search solutions soon.