|
|
|
|
||
|
|
|
My Journal |
||||||||
| Week One | ¤ | Week Two | ¤ | Week Three | ¤ | Week Four | ¤ | Week Five |
| Week Six | ¤ | Week Seven | ¤ | Week Eight | ¤ | Week Nine | ¤ | Week Ten/Eleven |
° ° ° ° ° ° ° ° ° ° ° °
Monday: When I came in this morning, I realized that the crontab Teng and I wrote last week ran on Saturday as planned. However, the programs it ran did not run correctly. Writing the BGP Table to the Database had an SQL Error that Jane was able to fix quickly. I also found that both the BGP Table and Relationship Files had differing dates depending on what was running. I hypothosized that this was because we ran the crontab at midnight and calculated the date at different points in the program. So at the beginning of the program it was still Saturday and at the end of the program is was now Sunday. So I rewrote some of the code so that the date is calculated only once and then passed to each program called. We then reran the crontab at 10:50am and it finished correctly at 4:15pm. While I was waiting for the crontab to run, I wrote another program to be added to the crontab which rewrites the applet's html page when the crontab starts to say that the database is being updated and then rewrite the html page again when it's finished to show the applet, with a heading stating the last time the database was updated. The general idea works, but I'm having trouble getting the time to be in the correct time zone.
° ° ° ° ° ° ° ° ° ° ° °
Tuesday: Today I finally figured out how to fix my timezone problem - turns out the timezone on my computer was set incorrectly and that was messing things up. So now the html page that is written when the database is updating states what time the updating starts and that it usually takes about 6hrs. Then Jane and I worked on an algorithm that would allow visualizing paths between any two non-tier one ASes. The general idea we came up with is running the non-tier-one to tier-one algorithm in both directions to all tier-one ASes. Then matching up the tier-one ASes of the first non-tier-one to the tier-one ASes of the other non-tier one and keeping those paths.
° ° ° ° ° ° ° ° ° ° ° °
Wednesday: This morning I finally solved the scroll bar problem in the applet. I was adding the toolbar after adding the panel and the size was getting messed up. So the applet is pretty much finished. Jane, Teng and I downloaded and set up a CVS system so that both Jane and I can work on the application without having two version of it changing independently of one another. To test it out, I added the right click menu optin the applet has to the application. After lunch, Jane and I continued working on the algorithm we started yesterday. We decided that a Breadth-First search would probably be better than the Depth-First seach we were doing, since we shouldn't have to look that deep to find a Tier-One. So we implemented a Queue and begain writing the BFS. However, we started getting a stack overflow error, which we traced to too many sql connections. So we rewrote both the applet and application to only connect to the database once and keep that connection open until the program closes, instead of having each new query be a new seperate connection.
° ° ° ° ° ° ° ° ° ° ° °
Thursday:Today I came in very early because I had to leave before lunch. Since Jane wasn't in yet, I couldn't continue work on the algorithm. So I added the color scheme the applet has to the application. I also fixed some menus to have default selections and made the node status frame hidden until something is drawn. Then I headed home and went to see Pat McGee in Boston!
° ° ° ° ° ° ° ° ° ° ° °
Friday: Day off, as I was still at home.
° ° ° ° ° ° ° ° ° ° ° °
Saturday: Went up to my friend's house in Maine to see some friends from Trinity.