>
Note: As Nick rightly points out in the comments, Dash is really a suite of related projects to help the community. The stats shown above are only a small part of that suite.
It looks like for the first time in a while that IBM may not have the most active committers at eclipse. It looks like the Individuals will over take them this year. Unfortunately I can’t compare this to prior years numbers as the site is very slow, but it is a good sign that we have so many individuals committing at eclipse. Personally, I take the DASH stats with a grain of salt. Mainly because it is very easy to manipulate the total Lines of Code metric. My own stats for the STAR Organization are probably over blown. I know I haven’t written that many lines of code. Most of that has to come from the work with the PsychoPath processor, and creating the necessary test suite based on the W3C test suite for XQuery. Much of that test suite code though I generated using XSLT and then tweaked as necessary.
DASH is one of those projects I don’t think the community takes advantage of enough. You can use DASH for all sorts of things.
- Who is a commiter on what project
- Is a person really a committer.
- How active of a committer are they from a code stand point.
- What projects do they have commit rights to.
- What companies are contributing committers.
- How active are those committers.
All of this is available through DASH.
Stats and Metrics are interesting, depending on how you sort and play with the numbers, they can either back up your stance, or be used against you. What does DASH say about you and your company? Or what can you make it say?
>Dash is a compendium of projects. You're just talking about the dashboard queries / commits explorer stuff. Dash also includes our beloved Portal and that oft-maligned Athena thing. Or maybe I have that backwards.
>@nick, so noted in the updated entry.
>Do you have some kind of auto-commit activated? One LOC per commit?
>@Sven: Nope, no auto commit function.
I do work though in a very iterative fashion. I commit my code in small chunks and with unit tests to go with it. As I said, some of these stats I take with a grain of salt. I don't know how they determine LOC and what is determined as a commit. So it's hard to judge if some of the stats beyond Active Committers are accurate.
>David,I just browsed through the dashboard (http://tiny.cc/HbYKX) and the second entry seems weird: 1 committer doing almost 300000 commits in 2009, each with an average of 1 LOC.Entries like that make me question the data as a whole. Besides, there are a couple of more entries with 5 digit commits and an average of 1 LOC, do people really commit such small portions on average?
>@Thomas: Unknown on the LOC and code per commit. It depends on how CVS is tracking that and how they are pulling that information out. If the data that the queries are getting is bad, or the algorithm used is bad, then who knows what it is actually saying.I know I tend to commit a lot (short small changes with unit tests to go with), but didn't think I had done nearly 300,000 commits.