Weeks 9 and 10

I got behind on the blog again.  We spent week 9 fixing small bugs in the software, but mostly working on the paper.  I ran lots of web logs and gathered data for the graphs.  Tara did most of the writing, I did some, but mostly edited and made sure wording was clear and concise.

We spent week 10 putting the finishing touches on both the software and the paper.  We read, edited, and re-read many times.  Hopefully we can get the paper published, because we sure put a lot of work into it.

Now I am sad because my new friends are leaving to go back home.  We have done a lot this summer.  At work we built software for holding bug competitions, held competitions, and overhauled the software so it would be more stable with many users.   Later, we held another competition, and our work payed off when the software worked properly.  We divided up, and worked on AI or Combinatorial testing approaches.  In combinatorial, we exceeded expectations, and got many different testing strategies implemented, then we put it in a graph for efficiency and continued to add new prioritization and reduction techniques, and we finished a paper which will hopefully be published.  Outside of work, we have gone hiking, swimming, rock climbing, camping, shooting, and river rafting.  We have had barbecues and potlucks, and have sampled many of the unique restaurants Logan has to offer.  In short, we have done so much together, and I will miss my new friends greatly.

Weeks 7 and 8

Since I am behind on this blog, and I cannot remember exactly which events happened in week 7 and which happened in 8, I am going to combine them.  Ever since the graph theory idea (the genius of Michael), we have been scrambling to get our goal of 2-way sequential prioritization and reduction and 3-way sequential prioritization and reduction working with the graph.  We also have needed to rewrite the existing 2-way and 3-way prioritizations and reductions, as well as the HGS algorithms.  When we had this partially done, we got an email from Renee requesting we merge our branch with another.  This proved impossible, because we changed how everything was stored.  We decided accomplish the "merge" by simply implementing the new features in our version.  So, we also got to write 2-way Intra (2-way pairs within a url) and 2-way consecutive prioritizations and reductions.  We finished writing all these new algorithms for the graph implementation about half way through week 8.  I was mainly in charge of doing the 2-way HGS algorithms (combinatorial, sequential, intra, and consecutive) but I also wrote a 1-way combinatorial reduction, 1-way HGS, and helped do some debugging on the HGS 3-way reductions.  The remainder of week 8 was spent debugging, and working on allowing prioritizations to be run on a previously reduced list of test suites.


Oh, I almost forgot, at the beginning of week 8 we did our summer trip.  We went to Jackson, Wyoming to go horseback riding and white water rafting.  Unfortunately, the horseback riding did not work out, because they don't take reservations, so instead we enjoyed a lovely lunch at some resort.  Most of us had buffalo, either burgers or steak.  So yummy.  Then we walked around Jackson, and went white water rafting.  It was so much fun!!  We learned team work, and bonded over our mutual clumsiness trying to get back into the boat after swimming.  Then we had a yummy steak dinner served by the rafting place.  The car ride back was interesting.  My car hit a deer, got lost, saw and almost hit 2 owls, 4 jackrabbits, and at least 20 more deer.  Good thing it was my own personal vehicle, which is kind of a junker to begin with, and that hitting the deer only caused one small dent (I won't bother fixing it) and ripped off my drivers mirror (now held on with duct tape until i have time to do a permanent fix).  All in all, we counted ourselves lucky, and everyone in my car is definitely bonded for life over those experiences.

Week Six

We continued working on reducing memory requirements and speeding up the test suite prioritization.  As we did this, Michael came up with the idea to store the tuples in a graph.  We estimated that this would use much less memory, so we spent the week creating the graph and re-writing the prioritization methods to use the graph.  I wrote a HGS reduction, which I will work on testing this week.  I also spent time fixing a problem with creating the test suites.

Week Five

We started working on test suite prioritization.  Dr Bryce asked us to add a 2 way sequential prioritization to some software and a reduction.  We finished this, did a 3 way sequential and its reduction, then found a bug in the existing 3 way combinatorial, and fixed this.  Unfortunately, this made it take a lot longer.  We figured out that the program was using a lot of unnecessary memory, and running out of heap space.  So, we began to rework the program to reduce memory requirements, and worked on speeding it up.

Week Four

We continued to make improvements to the bug testing competition website.  I continued work on the API, and converted a lot of files to use the new model.  Among these files were the files for chats, hints, and test results.  I converted these to write to the database instead of a file, then I wrote code to retrieve only new chats, results, and hints and display them to the user.  I also spent time debugging and testing.

Week Three

Monday we spent testing and fixing the bug testing software in preparation for the competitions.  Then, over Tuesday and Wednesday we held 4 competitions.  Sadly, there were several bugs that had not been worked out, and also our server had trouble handling the volume of traffic.  We spent the rest of the week reworking the system to make it more stable.  I helped write the API and worked on fixing small bugs.

Week Two

For week 2, we have continued making and testing the bug testing competition website.  We also changed the name of the site from Bug Stomper to Bug Catcher, so I redid the bug caught animation to show a butterfly being caught with a net to match the theme.  I also worked on the results box.  It previously appended admin hints the top, meaning the student would not see a new hint arrive unless they scrolled up.  We decided that we wanted the hints to come in order, so I created time stamps for the hints, test cases, and chats so they could be ordered.  Then, I added filters to the results box so that students can show/hide hints, chats, and all problem solutions (by default it only shows the test cases for the problem they are currently testing).  Other than this, I spent most of my time testing, finding bugs, and fixing them.