Greetings, ICS 413 students,
Some updates on Hackystat-related stuff:
* There have been a couple of reports of problems with the yesterday's release of the Hackystat sensors for Eclipse. The problem results from some development underway to add support for the Eclipse CDT (C development environment). We discovered that a dependency inadvertantly crept in this latest release that causes problems for people who do not have the CDT installed in Eclipse. We have just installed a new release (6.3.1105) on the public server which should fix this problem. I apologize for the inconvenience. Please let me know if you encounter any problems.
* On the server side, there are now a set of telemetry reports designed for ICS 413. After logging in to the public server, scroll down on the Analyses page to the "Telemetry Analysis" command. The reports are:
ICS413-DailyProcessStability
ICS413-DevelopmentProgress
ICS413-Review
ICS413-TestSizeAndCoverage
I suggest you select the "Chart" Report type, your ICS 413 project as the "Project", an interval consisting of the "Day's between today and around a week ago (or whenever you started work :-), and "**" as the "Template Values". Run through these analyses and think about the implications of them for your group work. Are there some members who are doing more development (i.e. Active Time) than others? Are tests and so forth being run daily? Is there work being accomplished on the Project most/all days? How about the coverage? Are files being committed regularly?
* If you are not seeing any data regarding Commits or Churn, the most likely reason (if you are indeed actually committing files to CVS) is that you have not set a Workspace Root corresponding to the CVS repository. Each person should login, go to the preferences page, and use the "Workspace Root Config" command to set "C:\cvs-repository\cvsnt\" as a Workspace Root. Once you've done that, Commit and Churn data should appear for your project. Note that your own commits and churn will not show up until you've set that Workspace Root correctly.
* When Hackystat receives UnitTest or Coverage data from you, the data contains the class name (such as edu.hawaii.413.graphster.FamilyTree) and the test or coverage data associated with it. However, project and telemetry analyses determine what data to look at based upon Workspaces, which as you know are a kind of platform and CM-independent directory path specification. So, for an analysis to use UnitTest data, for example, it must somehow translate a class, such as edu.hawaii.413.graphster.FamilyTree, to its corresponding workspace, such as graphster-xtc/src/edu/hawaii/413/graphster. You might be asking yourself, how does this bit of magic happen? The answer is that Hackystat uses FileMetric data (which is sent by the LOCC sensor) to build this bridge between Java class names and their corresponding directory path, because FileMetric data contains both the class name and the file name.
What this means for you is that you need to run the LOCC sensor regularly in order to provide Hackystat with the information about the classes in your system and the workspaces they correspond to. In particular, each time you create a new class in your system, Hackystat will not be able to map unit tests or coverage data you send to the server regarding that class to your project analyses until you run LOCC. I suggest that you run LOCC across your code a few times a week just to make sure everything is up to date. At the very least, everyone should make sure they run LOCC on the day before the weekly project review!
* On Monday, I will be reviewing your project progress following the outline in the project assignment. I will be looking in particular at the data from around this Wednesday (following the flood) to this next Sunday. You should be checking your Telemetry Analyses and making sure the data is both correct and indicates a smoothly running and effective development team.
Cheers, Philip
