The main reason I’ve never heard of this application is that this is a proprietary product which is sold for money. But now they released a new version which has a free version (free as in beer!) with limited features. So I gave it a test:
The installation went smooth – it was a shell script you had to run which asked you for the installation path and if you want to have symlinks to a system bin directory. It also tried to install some desktop icons which failed for my system, but that was easy to fix (the desktop files where directly in the main directory of the program).
After the installation is finished, you can launch xfriend with the double-click on the icon, and it fires up your preferred browser to point to
localhost:8888 – not the best address if there is something else running there already, but however, it worked for me. After a “first time setup” screen I got the main interface:
It is clean and handy, easy to use and self explaining – also, it is opened by the browser, so you do not have to start any additional software which is something I pretty much like! As you can see, there are some direct links to searches for specific file types, etc.
The configuration is also done over the web interface:
The search itself is as you would expect it: boolean expressions and so on:
The hits are marked yellow for the better overview, and a short excerpt is shown where a hit found a word. Additionally some tools are presented above every hit to work with these: you can open the file in a native program, save them to another place, or add them to “hitmarkx”: a basket of special search results:
You can add there all important and often needed search results to keep them together – a nice idea I must admit.
You can also search by different types of sources, like searching only the news:
Or searching only specific file types, like here *.odt files:
And, of course, xfriend can show previews:
Speaking about previews and supported file types: the list of files xfriend can index is quite impressive. Besides the usual stuff (MS Office, pdf, html) it can index odf files as well as exif headers and id3 tags. It fully supports ADS and NTFS/Share rights and can hand out a preview to every supported file type – when you purchase the full version, at least. The web page also claims that it can search in a more semantic way: considering the distance between words, and so on.
And xdot takes advantage of the fact that this is a client-server system: you can access the servers from the outside with ssl. Therefore a complete user management is integrated where you can create and manage users, make a detailed configuration for access rights and so on.
Even a web search is integrated: although it does not index visited web pages (as beagle can do with firefox and konqueror) it can search the web with overture (which is basically a yahoo company) and can present you the information found there.
In my rough test I did not run into any big problems – everything just worked. Sure, I would prefer an Open Source version, but this will probably not happen. The company could, however, improve their Linux version a bit by providing a LSB-rpm package (shouldn’t be a problem) and even better, could provide this in a XML package metadata repository to easily keep this software up2date.
But a real shortcomming is the e-mail support: you can configure pop3 accounts, but I would like to let the program know that I’ve stored my e-mails in maildirs or mboxes and that it displays everything found there with an e-mail viewer, and not with a text viewer (I’m not interested in the header!).
Other index plugins are missing as well, like a specific instant messenger index system, etc. That could be improved.
But, as usual, to get an own impression, just test it by yourself, you can download it for free-as-in-beer. And I think it is working surprisingly good, so it is worth a try.