My Weekly Journal

Home      About Me      My Mentor      The Research      Life In Seattle      My Journal      Final Report

Week 1 - This first week was a lot of setting up my workspace and absorbing as many research papers as possible.  I learned a lot about the RFID Ecosystem Project on the first day, as well as my access card and accounts.  After a couple of days, I established my project for the summer.  I then continued by starting my project proposal, which is my preliminary report at this time.  I finished off my week by playing around with coding in different Google applications in Java.

Week 2 - I started out the week by clarifying my proposal, the kinds of search queries my application would support, and learned about how to access the location data in our database.  This will enable me to use both location and human interactions as part of the search query constraints.  I also developed how the system will interact.

system_arch

In terms of coding, I at first attempted to make a Java Web GUI using Google Web Toolkit to make the interface and implement a Desktop Search.  It was working, but when I tried them out on Mozilla Firefox and Internet Explorer, it no longer worked.  I then moved on to trying to implement the same functionality with a Google Desktop Gadget.  I am having better luck implementing Desktop Search, however I need to learn more to have JavaScript and XML do a better job of interacting with each other to make my interface more dynamic.  At the end of the week, I had the opportunity to attend a talk from a visiting professor, Christoph Koch, from Cornell University.

Week 3 - At the beginning of the week, our group went on a hike to Rattlesnake Ridge, and ate out for dinner afterwards.  It was a great team building experience, as well as an enjoyable long weekend to refresh for work.  I have scenic pictures from the hike as well as the building I work in on my pictures page.  This week was a lot about coding.  I was fidgeting with my Google Desktop Gadget interface, as well as working on connecting with the database to get the needed time data to filter my search engine.  My gadget is at the point where it has filtering options to search with Google Desktop Search, and those results are displayed in a Google Desktop Search window.  My goals for next week are to have the ability to filter the data based on time.

Week 4 - This week has been all about filtering my Google desktop search results based off of time.  I have been requesting the XML code from the Google desktop page, and parsing it to get the information needed.  Over the course of the week, I have been able to successfully create a gadget that takes the input parameters of text, person, location (any combination, with text required) and output a search result page in html.  The page gives you links to the files, and the name of the file.  Unfortunately, my computer had some problems, so I lost some productivity for part of this week.  However, after dealing with bugs in my code and my computer, I had a somewhat good prototype.  The big goal for this next week is to make my searches much faster.

Week 5- I have been enabling users to select a particular meeting time, and refine the search results to be more accurate.  The search process was taking a long time, until I found out that I had a bad version of Google Desktop.  After I installed the new version, I could user their time filtering option in advanced search to receive my results.  I also coded so that the results page would be written to the user's temp directory to avoid unnecessary files on their computer.  In my search, I allow users to filter by file type.  However if one does cannot find the extension in the given options, I allow for them to filter by the extension with the "other" option.  At this point, my prototype was successful.  It is located below.

old prototype

Week 6 - This week, I ended up using Browse Timeline to get search results as opposed to Google Desktop Search alone.  This is because for example, if I created a file at noon, modified it at 2, and then at 4, I would only be able to see the file at 4pm in Google Desktop.  In Browse Timeline, you could view all 3 versions of this file.  Since people can edit the same file multiple times, I felt it necessary for the to be able to search on all modified times.  After parsing many HTML documents, the search engine is successful.  It works so much faster than with regular Google Desktop Search.  I have also started looking at more related work in preparation for the survey that is being deployed.  The purpose of this survey is to assess what laptop users do while they are at work.  For example, where they like to work, what objects or people are normally with them, etc.  Since I had successfully gotten results for a particular event, I decided to implement providing results for all possible events.

Also this week, we were doing test traces for the RFID Ecosystem database using typical work scenarios.  People would walk around with RFID tags while logging where they currently were to see if the database is accurate.

Week 7 - I continued trying to implement searching through all contingent meetings in search, however it was more challenging that I originally thought.  What it enabled me to do is to revamp my code.  I implemented old methods in more efficient and organized ways.  I had also deployed the survey to computer science and engineering grads, undergrads, faculty, and staff.  By the end of the week, I had added the feature to search through all meetings.  I had also created a new work event called "encounter event" for a variety of possible work events compatible with my search engine.

Week 8 - Based on the results from my survey I deployed, I made some improvements to my user interface.  I added the ability to filter work events by objects that they have.  In addition, I added the option to search for files modified Before, During, or After an event by a user specified amount of time.  I also gave users the option to configure their search engine so they could select any events they were interested in.  I have attached a screenshot of this process below.

options screen shot

More was done with the test traces for data logging analysis for the RFID Ecosystem database.  I also spent a significant amount of time this week preparing for my presentation.

Week 9 - On Monday, I gave a presentation on my GUI.  I have included a snapshot of my GUI below.

my gui

After Monday, I flew to Rhode Island for a conference at Brown University.  It was actually sponsored by CRA-W.  Here is a link to the Summer School.  It was a very enlightening experience, and highly recommend anyone else to do this if they have the opportunity.  While in Rhode Island, I continued working by preparing my machine for the User study that I hope to run my last week.

Week 10 - This week was spent debugging and preparing for the User study.  Unfortunately, due to lack of time, I wasn't able to get a successful amount of people for the controlled user study.  It was also a difficult task because people were required to use fictional data about their day at work to search for fictional files.  However, I am encouraged because in general people think this utility will be useful.  My end goal is to do a long term deployment of the system to see if it really is beneficial for users.  Event though my summer experience has ended, I plan to continue working to improve the GUI.