Blog Entries

August 2019
« May    

Increased Activity In KOffice

My interested has just been piqued by the following Tweet:

Commit rate in !koffice is insane; since 1-1 we already have 954 commits, compare that to the average of 480/month last year

So what did January 2009 look like compared to 2010? The following plots show commits and committers per day in 2009: … and here for 2010: We can see that the number of committers has increased and so this has clearly had a knock-on effect to the commits.  I have not run these numbers through R yet, but it looks like the increase in activity is proportionally greater than the increase in contributors. So, either the existing contributors are working faster, or are making smaller patches. Or both. Any other thoughts?

Be Sociable, Share!

5 comments to Increased Activity In KOffice

  • Albert Astals Cid

    Charts would have been easier to understand (imho) if they were one over the other and not two different graphs.

  • Jos

    or the commiters spend more time working on koffice

  • I agree with Albert Astals Cid, especially since you have different y axis.

    > I have not run these numbers through R yet, but it looks like the increase in activity is proportionally greater than the increase in contributors

    Maybe you can plot commits vs committers for January 2009/2010 in the same graph to compare.

  • d2kx

    Well, is nice and I am sure once it gets a proper UI it might get even more attention, but in the end it seems like it has its limits and is not as clean, innovative and nicely integrated as KOffice is, so I am greatly looking forward to KOffice replacing within the next months or years.

  • slangkamp

    Possible explanations:

    January 2009 was in the hard feature freeze for KOffice 2.0, while in January 2010 KOffice isn’t in the hard freeze for 2.2 yet. Bug hunting probably means smaller changes and fewer commits, than implementing features.

    There are commercial developers now, who didn’t commit to KOffice last year around this time. These developers can possibly work more on a certain piece of software than the volunteers.

    Different release cycles. In January 2009 KOffice was nearing the end of a three year release cycle. After that 2.1 was on a six months cycle and for 2.2 it’s six months again.

    The therefor the picked months are not the best choice, I think. It would be better to pick months from the same points inside the release cycle, so June/December, July/January and so on. The commits are not evenly distributed over KOffice, e.g. due to the involvement of the commercial developers areas like filters probably constribute many commits to the overall count (To verify that one would have to look how these components develop over time e.g how many commits there are in filters compared to the rest of KOffice).