1. One of the possibilities is to re-use the Google latitude smart phone app.
2. Use a GPS logger, such as SendLocation for iOS. For this option we need to implement a server-side script.
3. Create our own app for collecting on the smartphone the GPS data and send it to the server. We have several options. First, develop both one app for iOS and another for Android. Second, use phonegap for creating just one development for both platforms.

In any case we need to analyze the data in order to decide how to recollect it.

On the server side, we need to create a script that will be in charge either of received the data from the Smartphone and make it persistence in a database or to consume a web service (for example Google latitude) in order to access to the persistent data.

Server interface

We use jetty for create an embedded servlet container into our application. We use this to deploy a servlet that has the responsibility of handle the location data push to the server from the mobile clients.

1. The mobile clients construct a GET request that have the following Properties:
   • lat = latitude
   • lon = longitude
   • user = User ID
   • ts = timespan
2. The server respond to whether or not the location was saved into the database. This is implemented by adding a Header property into the HTTP respond, Result and the possible values could be either true or false. Also, there is a Header property in the case of an error, denoted as Error, with String values.

Server side

The idea is that we compute the desired Information Retrieval techniques as soon as a location data is sent from a user. So we compute the techniques per "place and user". In order to produce a better design for the Information retrieval processing, we can use the Observer Design Pattern in order to simplify the implementation of the Techniques.

Idea:

1. Subject (Servlet): shall provide support for attach, detach and notify to objects implementing the Interface Observer (DataObserver).
2. The Inface DataObserver defines the following methods: computeIRTech() with the data as parameter. The idea is that for each concrete class that implements DataObserver, we will implement the desired technique in the former method.

Regarding the timeline, you might want to think about how to organize the data before collecting them.

I don't think that you can finish categorization, search, filtering, and mining in one semester. We need to choose between the 4 possibilities.