Project 2 Feedback

Here are some of the problems I found in Project 2:

1. Some people are not following the Java Coding standard. Please make sure you review those guidelines. You will write code that other Java programmers can understand more easily if you format your code this way.

2. Make sure you are including *all* your leap data in a single file, and simply turning in a copy of the most recent version on each project. Some people did not have prior project data listed. This becomes important when you start doing historical analysis.

3. Many people are spending way too much time struggling with features (and sometimes bugs) in LOCC and Leap. If you don't understand something, and you've devoted 15 or 20 minutes to it, then stop and get help from Cam Moore (cmoore@hawaii.edu) for Leap and Joe Dane (jdane@hawaii.edu) for LOCC. They will be delighted to help you, and if you've found a bug, will fix it quickly.

4. Some people are not downloading the most recent release of the system. I know this is a pain for some of you, but it's a necessary evil. I know for a fact that some people have spent a lot of time struggling with a broken feature in LOCC or Leap because they thought they would "save" time by not downloading the new version where we've fixed the bug! One of the mandated course requirements was a reliable 56K connection to the Internet, and the reason I said that was because I knew you would need to download updates regularly during the semester. If you only have a 28K connection, it will take you twice as long as everyone else (and you will probably need to download Leap or LOCC at least a dozen more times this semester.) To succeed in this course, you must figure out a way to maintain a current version of the software.

5. Some people are answering their last postmortem question (how will you improve) in a very generic way: things like, "I will work harder", or "I will spend more time on design". Starting on the next project, I will grade you off if you respond to this question in such a generic fashion. It does not show enough thought and introspection on what you are doing. You need to come up with more specific "action items". For example, if you realize you are struggling with object oriented design, an action item might be "I will obtain a copy of Grady Booch's book on object oriented design and read it before the next project." Another one might be: "I realized that linear regression was a better way to map size to time than historical averages. On my next project, I will use that approach." Another one might be: "I realized that design patterns can simplify my design and implementation process. On the next project, I intend to use the Gang of Four book to determine if any design patterns apply to the requirements."

   I know that coming up with specific action items is much harder than generic responses, but they are also much more useful to you.

6. In your test case directory, most of you have supplied exactly what I asked for: a listing of the test files. Unfortunately, I realize in retrospect that this is not really what I wanted. :-) What I want is a listing of your test cases. The difference is that you can use a single test file to create multiple test cases which test out the behavior of your system in multiple contexts. This is a bug in my directions, not in your projects, so I'm just letting you know that I will fix this in the requirements for P3.