Long time no see

Today I've sent the report to my supervisor in Umeå, which means that the project is nearly finished.
All I have left to do here in Melbourne is to do a presentation next week at the Uni.

Furthermore, we have submitted a paper conerning the outcome of this thesis to the ACM mobility conference in Nice. Keep your fingers crossed....

Now I'm going to the barrier reef for a looooong vacation.

Over and out!

A demonstration

During the last weeks a lot of time has been spent on bug fixing and performance measurements. I'm also in progress of looking in to some user evaluations, which seems to be a bit more tricky than I thought, since I wan't to evalute the concept of application mobility more than the application's usability.

Any way, here comes a short low quality video demonstration of the Mobile YouTube Player.



As you can see the applications dimensions are adapted to the new device, as well as the video quality is changed from HD to standard quality when the application is moved to the Sony Vaio. This is a bit harder to see though. =)

Bad system design?

Today I've spent three hours drawing the arrows in following class diagram without having them to cross each other too many times.... I wonder if that might be an indication of bad system design? (Microsoft Visio's fantastic "self bending arrow functionality" didn't exactly help either")





























However, despite the messy class diagram the Migration Agent is starting to work rather good, with functionality like multicasting, distributed server solution, p2p networking, video quality adaptation, and so on :)
A video showing a migration is to come...

Over and out

System Design

During the last weeks I've been working with the system design, a lot of UML diagrams such as use case diagrams, sequence diagrams, class diagrams, activity diagrams have been drawn.

To my surprise I come to realize that I'm now one week a head of schedule, which means that I've had time to start some implementation, the outcome so far is the Mobile YouTube Player, that actually works fine except for some exceptions that are thrown every once in a while.
The Mobile YouTube Player is basically the YouTube's video player wrapped with a java shell that makes it possible to save and resume the player's execution state.

The next step will be to build the "Migration Agent" that will take care of moving the Mobile YouTube player between devices.

Furthermore I've started to plan some user evaluations of the system. Anyone knowing a good tool for evaluating a system's functionality and not its GUI??

Cheeers!!

Working hard

While the Aussies are enjoying their Labour Day, I'm keeping myself busy with creating Use case scenarios and drawing UML diagrams. A man got to do what a man got to do!

System Requirements

I've now spent two weeks in order to specify tools that will be needed for the implementation. My findings are:
(1) YouTube doesn't make it easy for people who wan't to control their videos without using their video player.
(2) Java Media Framework is the most annoying thing ever buit. Mainly since it doesn's support any of those formats that are most common today.

At the moment I think that the best solution would be to use YouTube's own player, since it got all functionallity needed. However, since the mobility part will be built in Java, I will need to find a way to communicate between Flash and Java. hmmm. any ideas?

At least I have a plan of how to build the "Migration Agent", whos function will be to indentify and locate the user as well as suspend, resume and migrate the video application.

On Monday I will initiate phase 4 of my thesis, which is system design.

Finally Friday!
Cheers!

Two weeks of article readings

I've now finnished the first two weeks of my in depth study. The outcome is a litterature survey in application mobility. Finally I get to put away the articles for about 12 weeks before I will finish the in depth study!

The next step will to identify and specify system requirements for my Implementation. The following scenario shows how the system basically will work.

I think that is all I have to say for now. Might just post a picture of the view from the floor where I'm working to make this blog a bit more fun. Furthermore I just want to add that the air conditioning in this building is extremly effective, my hands are so cold that I can't maintain my usual typing rate on the keyboard.

Research/Work Plan

After a few meetings I've now managed to write a preliminary research/work plan which goes like:
----------------------------------------------------------------------------------------------
The research plan will be divided into mainly three parts:

1. Mobility. The first part will concern mobility of an application. This means more concrete
that I will as a start build a Java application that is able to stream a YouTube video. The
second step will be to move this application from one computer to another on a given
command. The computers will use a wireless peer-to-peer network in order to migrate the
application. A server included in the peer-to-peer network might also be used to avoid
scenarios where the sender device is turned off before the receiver device requests the
application. Furthermore experiments concerning if the buffer should follow with the
application during migration, or be downloaded when the application is resumed is needed
since this will probably influence on how seamless the migration will appear.

2. RFID. The second part will be to use RFID-technology to discover and identify a user. By using
passive RFID-tags on the users and RFID-readers on the devices, the devices will know if a
user is present and who he/she is. This will be used to make the mobile application usercentric,
which implies that the application will follow the user, hence run on the device
closest to him/her.

3. Migration between different devices. The third step will be to move the application
between one laptop computer and one smart phone. This step will also involve adaption of
the GUI which is needed because of the significant differences in the interaction possibilities
between the two devices. The application must therefore be able to modify its appearance at
runtime. This can be achieved by using SwiXml [1] which is a GUI generating engine for Java
applications where GUIs are described as XML documents that are parsed at runtime.

Furthermore two extra parts might be added, depending on how time consuming the first three
steps will be.

4. Step four will concern identification of new relevant context that can be used to enhance the
applications performance

5. The last step will be to try the implemented prototype on the M4 project.
----------------------------------------------------------------------------------------------

That's all for today. Now I'm going to IKEA, Over and out!

First day at work

To day I will get started with the in dept study of my thesis work. Hence I got some nice article readings to look forward to. As a start I will look at related work within the area.