This is just spectacularly cool. :-)

To start, how about a regular Hackystat Project-level analysis that lists each workspace, the number of in-dependencies, and the number of out-dependencies. Make each column sortable by clicking on the column header (see the Adoption command for an example of how to do this).

From this command, we can immediately see workspaces (i.e. packages) that depend upon a
lot of other packages, and the packages that a lot of other packages depend on.

Also, if you could send me an XML version, I can make it available to the 413 folks as input to their graphster program. That will allow us to see a graphical network layout.

Cheers,
Philip

--On Friday, November 19, 2004 1:17 PM -1000 Aaron Akihisa Kagawa <[EMAIL PROTECTED]> wrote:

Hey All,

We have a new report and sensor for Dependency Metrics. I've added the sensor 
to the
daily build and we should be getting data every night.  You can see the report 
at
<http://hackystat.org/hackyDevSite/build_log/20041119/Hackystat-UH/dependency/package2p
ackage/index.html>.

One of the interesting packages in this report is org.hackystat.kernel.user.  
There are
more than one hundred other packages that use this package. This type of 
information
identifies some interesting things about quality.  For example, changes to the 
user
package will definitely highly impact other code.  I hope to use Dependency 
information
for the Priority Ranking Software Review.

Currently, I've implemented a DailyProjectData class for this metric. However, 
I didn't
create any reducers and/or analyses because I'm a little unsure how to provide
something useful.

thanks, aaron







Reply via email to