A blog for the enterprise open source community
A totally unscientific look at the historical trends regarding forksMatthew Aslett, June 20, 2011 @ 1:13 pm ET
Back in September last year, as the LibreOffice announced its separation from OpenOffice.org and Oracle, I published a quick post comparing the relative success of a variety of forks.
Inspired by Alan Bell, I used Google Trends charts to illustrate the relative level of interest in a number of projects and their forks. Hardly a scientific study… but Google Trends charts do provide us with an indication of the level of interest in a fork, compared to its parent, illustrating the impetus a project has to separate from its parent.
Brian Proffitt published an interesting post last week using Newtonian physics to describe how a fork needs enough velocity to escape the orbit of its parent.
A look at the charts used in my previous post shows how quickly a child can overtake its parent – at least in terms of Google search results – if it has enough momentum.
It took Joomla (in red below) just two quarters to overtake Mambo (in blue) , for example, while it took Ubuntu about a year to surpass Debian.
Other forks are less immediately successful. It took Adempiere about three years to overtake Compiere, for example, while vTiger (in red below) has taken six years to close the gap on (but not surpass) SugarCRM (blue).
Brian’s post argued that the history of forks did not favour OpenOffice.org, but there are other examples of forks that have failed to achieve the success of ECGS, X.org and Webkit (Brian’s examples).
Perhaps this is not the fault of the child, but the nature of the parent, however. Dave Neary suggested that vendor-driven projects have an advantage over community-driven projects in artificially maintaining momentum in the face of a community fork.
That would seem to be a reasonable suggestion, but it does beg the question about what the when, or if, LibreOffice might gain enough momentum to overtake OpenOffice.org.
For one thing, OpenOffice.org does not fit the description of Debian, or SugarCRM. It was a vendor-led project, but does not remain so (despite IBM’s likely dominant role in its future).
For another, the LibreOffice fork came from within, rather than being the result of external influence on the developer community.
So will the OpenOffice-LibreOffice chart more closely resemble Mambo-Joomla or SugarCRM-vTiger?
Almost three quarters after the split, LibreOffice (in red) has certainly narrowed the gap on OpenOffice.org (in blue), but not by a significant margin. However, it is too early to draw any conclusions.
Add Comment Categories: Software