> Also, look at the downloads: > http://people.apache.org/~vgritsenko/stats/projects/axis2.html#Downloads-N1008F > vs > http://people.apache.org/~vgritsenko/stats/projects/cxf.html#Downloads-N1008F > More downloads <=> more usage <=> more scenarios tested <=> more bugs found!!! > Less open issues != better software
I agree that looking at issue statistics alone is not enough to evaluate the state of a project. However, download statistics are even more unreliable. Here are the reasons: * AFAIK, all use cases of CXF are based on either embedding it into a Web application or deploying it as bundles into an OSGi container or using it as a library in a stand-alone application. On the other hand, CXF has no equivalent to the container-like deployment scenario that Axis2 has with its axis2.war or its stand-alone server. That means that for people using Maven (or equivalent) there is no reason to ever download a CXF binary distribution. This is different for Axis2, where people would still download the binary or WAR distribution, even if they use Maven to build their services. * The normal way to download a distribution is via a mirror site. I don't know how the download statistics have been compiled, but I doubt that the mirror sites forward their access logs to the ASF for aggregation. * As the footer of the statistics page says, "In no way these charts should be considered complete or official data." Since the statistics primarily focus on Web page access (for which there are no mirrors) this in particular applies to the download stats, which are probably only a by-product of the analysis. Andreas --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@axis.apache.org For additional commands, e-mail: java-user-h...@axis.apache.org
