< Insert Your Favourite > Desktop Search Hackfest

Good news - a (read: the first ever) Desktop Search Hackfest is being
planned after the Maemo Summit, in Berlin.
http://wiki.maemo.org/Desktop_Search_Hackfest

Not so good news - I will not be able to attend. It's not easy to
sneak out as a grad student from an US university (bonus if you are an
_alien_). Joe is not going either. Our in-house Xesam guru is about a
join a real job, the money-paying kind. It might be hard for him to
attend either. I dearly wish Beagle could somehow benefit from this
meeting. Good luck to the other projects. Get some work done and make
the users happy. It is nice to see that Strigi/Nepomuk devs are
attending. And thanks to Nokia for making this happen.

On a side note, look at the number of participants from the Tracker
project ! Wow, they're booming. Double kudos to them.

Creating new backend - help wanted

Three years ago this month, the above was the subject of my first post to the
dashboard-hackers mailing list. I was writing a backend for Akregator. Its
immaterial what Akregator is; I was mostly pattern matching back then and
that is how I got familiar with C#.

I remember submitting a patch to beagle bugzilla a few days before that. It
was to filter JPEG JFIF comments because that is where digikam stored all the
descriptions. Now it uses IPTC tags. Does not really matter since beagle
indexes both.

Backtracking further, I got interested in beagle several months before that. I
had this incredible urge to add /author/, /title/, /subject/ tags to the PDF
research papers that I download and use beagle to search among them. The
first time I visited the beagle website and downloaded the tarball, the huge
list of dependencies really really scared me off. I used to allow more Gnome
on my desktop back then than I do now but nevertheless building beagle seemed
like a daunting task. I later freed beagle from many of its dependencies; I
feel that is my biggest contribution to this project.

I made a second attempt later; soon realized that even though I can get beagle
going, there was (is ?) no easy way to modify PDF document properties (in
linux). Then I switched my goal to index my pictures, which I have started
tagging and adding comments to, using my new found love digikam. Thankfully
the one-liner JFIF patch did not require much C# knowledge. It would not have
mattered anyway, C# is darn easy to pattern match.

After all these months, I still haven't managed to build my repository of well
tagged PDF papers (if I wait for several more months and stop wasting time
then the need would be gone forever). I do keep a static index for my
pictures but rarely search them. What a waste of time :-). Silly me!

As I am about to sign off this blog entry, I noticed that I was siging off
as "Bera" in my initial few emails. I wonder when did I switch to
my "signature sign" dBera.

Did I mention that beagle 0.3.8 was released some 48 hours ago. Go go get it.