I’m amazed that Ubuntu doesn’t have a good default Desktop Search tool. I’ve been using Ubuntu for a while now and the lack of a perfect desktop search is felt very strongly.
The need for such a tool is greater in Linux than in Windows considering you have a greater degree of freedom in configuring obscure files to make the OS perform the way you want it to. From that perspective, it has become quite a pain searching for files of something I had installed a long time ago. Also, I’ve been reading a few research papers of late and I’ve been making some notes in txt. ReFinding something is a nightmare if you can’t remember the filename.
So I started looking into the different desktop search options I had.
Beagle is widely considered to be the leading tool for desktop search on Ubuntu. Its written in C# and runs using Mono. It uses Lucene’s C# implementation for indexing and it seems to be good. It was launched with (a bit of) fan-fare back at the 2004 Guadec Conference in Norway. It is quite similar to Apple Searchlight and was, interestingly, announced on the same day as Searchlight. Norway being 6+ hours ahead of the United States, it was technically released earlier!
But Beagle faced performance issues regarding its ‘crawling’ process which was very processor heavy. Users grew irritated with the constant processing involved in keeping a fresh index that many disabled the tool. Some time ago development on the Beagle Project was stopped and so, I believe, did the adoption by new users.
As a product, though, Beagle is quite good. It has a rich feature set and if you ignore the index refresh issues, its quite good at retrieval. It indexes all imaginable file types and has some interesting features, like the support of glob expressions which I haven’t found in other search engines (except Recoll).
Tracker is short for Meta Tracker. It is a light and fast desktop search implemented using C. Tracker’s lightness has meant that it doesn’t index the innards of some types of files – pdf, ppt, mailboxes etc. But it is faster and doesn’t eat up processor cycles, so its a more quick and dirty solution. And of course if you can’t find something using it, you can always switch over to good ol’
Earlier, documentation was very poor on Tracker, but it has improved, I feel.
Google Desktop Search
Google Desktop Search isn’t open source, so it doesn’t get brownie points. But it is quite useful. It indexes almost all file-types including pdf and ppt. The engineering behind it is good and it doesn’t hog (too much) processor time. If you can get over the fact that Google will get even more of your information, then this is the desktop search tool for you. I prefer to have at least a bit of my own life to myself, so I’m left with very few options.
There are many other tools which I haven’t mentioned here.