Google Desktop for Linux: Plain ol’ desktop search

Linux users finally have their own edition of Google Desktop. The beta release was announced Wednesday, and I’ve been putting the application through the wringer since then to see how well it stacks up on Linux. I found that it’s a nice offering, but it slows a system down noticeably.

While Google has finally delivered a Linux version, Linux users still aren’t getting the red carpet treatment. The Windows and Mac OS X pages for Google Desktop show some features missing from the Linux version. For example, Google Desktop for Windows has a sidebar with Gadgets and Plugins and an enterprise edition, while Linux gets … just plain ol’ desktop search.

Installing Google Desktop

Installing Google Desktop is a snap. I grabbed the Debian package and installed it on my Ubuntu 7.04 system, even though the only “supported” version of Ubuntu listed is 6.10. After being installed, Google Desktop immediately started going through files and building its index of what’s contained in local files, Web history, and so forth.

Indexing took approximately forever on a system with an AMD Athlon XP 2600+ with 1GB of RAM. I configured Google Desktop to search all of the default directories — my home directory with about 1.2GB of files, /usr/share/info, /usr/man, /usr/share/man, /usr/local/share/man/, /usr/local/man, and /usr/X11R6/man — as well as an external USB disk with about 75GB of data.

When I began indexing, the program gave me an estimate of 5.5 “idle” hours to complete. Google Desktop’s FAQ says that indexing occurs when the computer is idle, but I also noticed a slight performance hit while I was working — even saving files from Vim took a few seconds. I suspect the performance hit is less noticeable on more recent computers. You can pause indexing at any time if you’re seeing performance problems, or for any other reason.

More than two hours later, the Index Status page reported about 5 hours left to complete. That’s the point where I went to bed and hoped Google Desktop would churn away at my data while I slept. It did, but it was still churning when I got up. In all, it took more than 15 hours to run.

The Index Status page is nice — it provides an estimate of how much of your data has been scanned, and how many files are in the index, including a breakdown by types of content. For instance, you’ll see how many email messages, files, media files, and how much of your Web history has been scanned. It also shows the time of the latest item scanned.

Using Google Desktop

Once Google Desktop is running, you can search by using the Google Desktop page in your browser, or you can bring up a search box by pressing Ctrl twice in succession. I like that shortcut — it’s easy to remember, and doesn’t conflict with existing desktop shortcuts.

After you enter a search term, the results show up as a Google Desktop page in Firefox looking much like the results of an online Google search. Google Desktop integrates the browser links with the desktop, so when your search term matches, say, files on disk, you not only get an “open file” link, you also get an open folder link, which opens the folder in Nautilus under GNOME or Konqueror in KDE.

New files, Web pages, and other documents get added to the index immediately. If you make a change to a file, the new version shows up nearly immediately in searches.

Not only is information added to the index quickly, Google Desktop also caches data — which means you can retrieve deleted files using Google Desktop. I tried this by deleting an HTML file with a few keywords and then searching for those keywords. Sure enough, the file turned up in the index and I could open it via Google Desktop, even though it was gone from the filesystem itself.

This behavior has positive and negative implications. If you’ve accidentally deleted a file, you might just be able to recover it using Google Desktop — hooray! On the other hand, if you deliberately deleted a file, someone else might be able to recover it using Google Desktop.

“Advanced features”

Google Desktop comes with an option to turn on so-called Advanced Features. The name is a bit misleading, because one expects that “advanced” features would include user-related features. Instead, with Google Desktop, it means that Google will glean some “non-personal” information from your computer, including “summary information, such as the number of searches you do, the time it takes for you to see your results, and application reports we’ll use to improve the program.”

This is an option you see when you start Google Desktop, and I’m guessing many users turn it on thinking, “Oh, boy! Advanced Features, gimme some of those!” never realizing that they’re really just giving Google permission to send data back to the Google mothership. I have no problem with Google including an option to collect data, but this phrasing strikes me as a sneaky way to get users to participate without understanding what they’re really doing.

Google Desktop vs. Beagle

While the Google folks have been toiling away at the Linux version of Google Desktop, Linux users already have a serviceable desktop search program in Beagle.

Both Beagle and Google Desktop index browser history (if you use the Beagle extension) and local files of many types. Both do a good job of indexing data and making it easy to find files. They both seem to miss some files too — for instance, neither Beagle nor Google Desktop seem to have indexed my chat logs from Pidgin or XChat. I searched for terms that were in my chat logs, and got back nada. In general, I haven’t noticed much discrepancy between results returned by Beagle and Google Desktop.

It does seem to me that Google Desktop puts a little more strain on my system. I haven’t noticed system slowdowns while running Beagle in the past year.

Google Desktop does have a clear advantage when it comes to Gmail. If you have a Gmail account, you can include Gmail messages in your desktop search — which means not having to go to Gmail, or download mail from Gmail, to be able to search your mail.

Google Desktop is still in beta, so its performance will probably improve before its official release. If you’re, um, searching for desktop search, download it and give it a go.