Power of Wikipedia: Italian Wiki Shuts down to protest against wiretap

This shows the power of the wikipedia, Yesterday I was working on putting together some research questions to the project that I plan to work with Paolo Massa and checked my wikipedia user page on italian wiki. First I didn’t understand what was written so I translated and I realized that Italian Wiki has started to protest against the new wire tap act. This shows the power of open collaboration and especially in wikipedia.

Everyone at WikiSym loving the WP:Clubhouse? An Exploration of Wikipedia’s Gender Imbalance

I am currently attending WikiSym Conference and its great to see that all the hard work that we put on the gender paper getting rewarded. First we won the best paper award and during the paper presentation audience had great questions along with compliments for the work that we carried out.

If you are interest in our paper and haven’t read it here you go : WP:Clubhouse? An Exploration of Wikipedia’s Gender Imbalance

Gender Gap out of the door, now going over ideas basket

So we submitted the paper to wikisym, I had so much fun while working on the Gender Research paper. As a data junky slicing and dicing the data and sending them over many analytical processes was astonishingly thrilling. We came up with some great research questions and hypothesis it was fantastic to see how our hypothesizes tally with the results we got from our data. But to find out you need to read our paper and will put a link here sometime soon.

Now, I am going over my idea basket to find out which thought I should present to my advisor to get his thoughts on my next research angel.

Developer says: Editors cant live with them cant live without them

I just switched myself to JEdit, after carefully analyzing I decided to go with JEdit simply because its free. I really like TextMate but its student version is $50 bucks so I decided to go with JEdit. And boy it did fascinate me, simply because I can load remote files and simple work on them. without having to SCP them to my account.

Gender Related Topics are Facinating Research topics

Gender gap is an interesting topic and it gets more interesting when you narrow it down to different domains. Since women have a extra-chromosome they should be the dominating gender in the world but in most areas that’s not the case. I am currently working on research publication and things that we found blew my mind off. Some of our research questions turned out to be way different that we predicted. I ll promise to put more information once the paper get publish. Till then stay tuned.

GWT and Multiple Permutations of Compilation

While I was poking around GWT and eclipse I realized that GWT compiler carries out multiple permutations of compilation process. My initial thought was, may be they are doing it to support cross platform compilation. This doesn’t happend often but this time my guess was spot on. So, I did little research and found that GWT doing its multiple permutations of compilation to support all browsers and this can be changed and you have the full control of it.

I found some developers have complained that enterprise level development take longer compilation and that has become an issue among many developers where length of debugging has increased. Just to show what exactly I am talking about, I have copy pasted the simple compilation of GWT MyWebApp Application.

But there are ways of working around. You can reduce the compilation just to compile for one browser. In my opinion is this can be done by using package/modularization (java package structure) with the help of ant to do the build process. But this article is not to talk about use of ant of java package structure.

By default GWT get compiled to support following browsers.

ie6
ie8
gecko
gecko1_8
safari
opera

Do we really need all these browser during the development time. I think thats a personal call but when I develop something I hate to wait the compiler to run for super long time. If thats the case for all of you here is the solution. I would also like to bring up another aspect on reducing compile time during development. Imagine our small application (MyWebApp) is not as small as we think and we support multiple locals. In this case to support all the browsers we have to compile the application in the default mode. Just to make this argument lets say we support 10 locals.

Here is the calculation for compile time: 36.654s * 10 = 366.54s this is fine when a build master is doing production build. But as a developer I would like to do a quick build and see my changes.

this is the file that you need to change to get different permutations.

this file come in handy if you want to separate the project into multiple modules or rename the module to make the module name shorter. To reduce the number of permutations or to build the application just for one browser, we need to tell the compiler to do so by using this file.

Module xml file(s) are located on the java class path and they are always referred by their logical names. So let say you are going to develop an application call MyWebApp, and you own the “boo” domain then you can find this xml file in com.boo.mywebapp under the name of MyWebApp.gwt.xml

by default you have set it up to compile for all the browser and it should look like this or something similar to this.

To reduce the compiler to do only one permutation you just have to pick a browser that you would like to do the development and add that as a property value, in this case we use the user.agent property and set the value to gecko1_8 which is firefox

here is the *.gwt.xml file looks like after adding the property value.

Lets see if this really reduced the compilation time.

Sure it did, and you can also see that it only did one permutation. Similar approach can be taken to locale where you define the local you want to compile with.

Dimension to analyze strength of the relationship

As a grad student I read many publications, and as requirement for one of my grad class, I read a paper published by University of Chicago urbana-champaign. The title of the paper was Predicting Tie Strength With Social Media . It was very interesting and I was fascinated by how social media can use (to some extend) to predict the strength of the relationship. But I do have my opinion on this paper and here it is,

Finding a dimension to analyze strength of a relationship is not an easy task. Human behavior and human thought process is very complicated. Emotions and feelings are two things that changes in a rapid pace. It would have been nice to identify the status of the users mood while they were answering because that can have an effect on emotions and feeling which may change the strength of the relationship.

Until the results sections I was questioning this paper with respect to how would they determine a relationship between a couple, most interestingly a relationship between husband and wife. They could very well open Facebook account and never communicate via Facebook. But results sections really answered my questions because they identify outliers seems to have higher strength in the relationship. I think this is a reasonable argument where I find its very true. It did fascinate me how these outliers very much could represent the married, engaged or people who are in a relationship. So can we identify relationship status based on strength of the relationship between each other.

I don’t agree how they used education as a social distance variable. I don’t think that can be used as measurement of the relationship. In real world people don’t become friends with each other by asking level of person’s education. Prime example for this is Asymmetric Friendship that they have identified. Most grad students end up having great friendship with their advisor and continue through life time. Now they may end up with a PhD later in their life upon graduation but not necessarily.

But overal this paper was really interesting. It also reminded me the The Social Network movie. According to the movie, Facebook was started based by ranking pictures of friends. This research is not that far from it.

Research having a blur phase

Its about time I pickup my graduate research and get on with it. After my last two papers got reject I decided to take a little time off, but I don’t think taking time of from that is the write thing to do. Yes, I am not in the same state of mind as I was in couple of months ago but I think going back and pickup up my work will help me to avoid thinking about many things that I can’t control.

Proposal got approved

Just got the news that my proposal got approved and it’s time to get my implementation hat on. Hopefully I will be able to do a publication based on this proposal by sometime in January. The implementation will be based on Couple of papers that I read at the end of the summer. Proposal was written targeting a paper written on Fractal Dimension at its usage for better clustering. I will implement the propose algorithm for better clustering based Fractal Dimensions. I plan to keep you update with the progress.

Hosting Facebook Data Mining Project in Google Code

I have created a Google Code project to host our Facebook base Data Mining development. We will be checking in all the development starting from extraction to all the development of our mining concepts. The project name is data-extraction-facebook. Since this is a research project, only myself and Mark are allowed to do check-ins. All the development will be done by myself and Mark and Anjana will provide her weka knowledge to compare results that we get from our data mining implementation with Weka.