Paper blog

For a relatively long time I have been interested in the Mobile Application Development world. I think that the world we currently live in is rapidly becoming a friendlier place for both people who develop these products and end-users. It's a vicious cycle, a "build-it-and-they-will-come" type of situation. Today, good business practice translates to a smooth transition between whatever it is the business does and the technology it uses to carry it out. One of the many pieces of this technology puzzle is Mobile Development.

I chose a paper written by the Department of Computer Sciences at the University of Wisconsin Madison dealing with the analysis of mobile users' experience through two popular but very different applications. In order to measure and analyze the differences in user-experience they built a measurement framework and library, which when installed on applications could measure anything from "battery drain rate" while using the app to session length.

I think this way of "behind-the-scenes" data-analyzing is very effective for the developer given that they can capture instantly the way in which users respond to any particular feature. Additionaly, it is very helpful because any bugs or feature issues can be addressed right away when it is first detected. I do think it is somewhat intrusive on the end-user given they get access to a lot of their information but I imagine they agree to the terms when they decide to download the application. One of the most interesting facts I learned is that devices with relatively same CPU, memory, and battery capacity showed considerable differences in battery performance. They mention the Kindle Fire example where controlling the screen brightness reduced the average battery drain by 40%. I imagine this allows developers to engineer different enhancements to the application, corresponding to the device it is being downloaded in, so that battery can last the most in each one. Different devices may demand different things and, although painful for the developer to create different versions of the same product, this can help end-users some trouble.

Another cool aspect they mention on the paper is the fact that usage behavior is impacted largely by its platform and audience. For educational applications, for example, the daily active users was highly correlated with the day of the week. This may come into play in significant ways for the developers. If the application is being used mostly on week days (for schools or universities) then there is a clear need for their servers and everything supporting the application to be at its best performance during those days, allowing them to perform any routine tests, updates, or experiments on weekends. These analysis definitely shape the way an application interacts with end-users.

Altogether I think this is a very neat way of capturing data, in the paper they break down all the sections that were analyzed and the corresponding information. I think that this type of analysis saves developers a whole lot of surveys and marketing strategies to determine whether they are succeeding in the way they are portraying their product. It is very efficient and time-saving. I do consider that Mobile Development will continue to be an outstanding piece of our technology environment and that they will continue to come up with ways in which developers measure our response to their applications.