I'm set to learn Qt. I stumbled upon it while searching new perspective to my Netflix Prize solution. I found Icefox's Framework, which is coded with Qt.
I got interested in it. The simple makefiles, quick build process and, for now at least, decent looking framework in general. Works both in Windows and Linux apparently, neat. I've worked with GTK+ earlier, and while I liked it, I also hated it. Coding in C++ wannabe C, sigh. Autotools, sigh.
Lately, Nokia bought Trolltech, the main developing company of Qt and Qtopia. Knowing that Nokia has put a lot of balls and cash behind Symbian, it should be interesting if they will shift their focus to Qt and put all their weight behind it. Sooner or later (my guess is later), it will happen. Symbian is not a system that will live much longer, its such horrible system to code for.
I'm just starting my journey with Qt, but I have high expectations. I'll later comment how it went, and possibly share my wtf-ish pieces of code with you. Maybe I'll write up a tutorial from moving from GTK+ to Qt.
Showing posts with label netflix. Show all posts
Showing posts with label netflix. Show all posts
Friday, March 28, 2008
Wednesday, November 28, 2007
Lack of updates
I probably don't have any readers for this blog, as I haven't announced it anywhere (well, maybe the Facebook notes auto-import thingie counts as an announce?), so there is no-one waiting for my updates on this blog. Its been well over a month now, and I try to sum it up now.
Lets start with the Facebook. Since I joined it, I spent half a day finding my friends (more or less successfully) and configuring the glittery apps the place is full of. As anyone who uses Facebook knows, the apps seem to be "the salt" of Facebook, and they are true. I wouldn't be using Facebook nearly as much if there wasn't any.
However, there already is too many apps; the same functionalities exist in many similar apps and some users finding one apps before the other, they spread that one to their friends and they pass it to their friends and so on. Eventually you get invited to the "same app" hiding in different skin a gazillion time by different friends. And oh boy, is that annoying!
Well, in my book Facebook gets a B- for a good try. We'll see what other sites the future brings.
Another thing I've spent a bit of my time is the Netflix Prize competition. In short, they are offering one million dollars to first team to improve their movie recommendation system accuracy by 10%. It has been running for a year now and team closest to the prize has achieved about 8.5% improvement.
Realistically, I don't believe I can beat even that, but I'm using the competition as a framework for my Artificial Intelligence software project at the university. Getting study credits for something I have personal interest is a great thing! Another interesting thing about the competition is that you get an "authentic" dataset, with 18000 movie titles, 500000 users and 100 million ratings. That is hard to get by and registering to the competition only to download the dataset is worth it.
Thats all for now, more updates later!
Lets start with the Facebook. Since I joined it, I spent half a day finding my friends (more or less successfully) and configuring the glittery apps the place is full of. As anyone who uses Facebook knows, the apps seem to be "the salt" of Facebook, and they are true. I wouldn't be using Facebook nearly as much if there wasn't any.
However, there already is too many apps; the same functionalities exist in many similar apps and some users finding one apps before the other, they spread that one to their friends and they pass it to their friends and so on. Eventually you get invited to the "same app" hiding in different skin a gazillion time by different friends. And oh boy, is that annoying!
Well, in my book Facebook gets a B- for a good try. We'll see what other sites the future brings.
Another thing I've spent a bit of my time is the Netflix Prize competition. In short, they are offering one million dollars to first team to improve their movie recommendation system accuracy by 10%. It has been running for a year now and team closest to the prize has achieved about 8.5% improvement.
Realistically, I don't believe I can beat even that, but I'm using the competition as a framework for my Artificial Intelligence software project at the university. Getting study credits for something I have personal interest is a great thing! Another interesting thing about the competition is that you get an "authentic" dataset, with 18000 movie titles, 500000 users and 100 million ratings. That is hard to get by and registering to the competition only to download the dataset is worth it.
Thats all for now, more updates later!
Subscribe to:
Posts (Atom)