Today I had a great reminder of why I love open source software and why I spend a bit of my time contributing R code to several packages.
In the vegan package, I had included some code to do an analysis of multivariate dispersions. Bone-headedness on my part meant that if you tried to plot the ordination results behind the model/test for any axes other than the first two, the code failed. A user (Sarah Goslee) spotted this and emailed me with a fix to the code, that I had integrated and added to the vegan source tree on R-Forge within about 30 mins of Sarah's email landing in my in-box.
This particular bug was fixed in the forthcoming 1.17.5 release of vegan.
Don't you just love open source!