Turns out, there are some anomalies. The data for 2.5.x are probably incomplete due to the changeover to Launchpad several years ago. Also, Plone 3.2.1 shows almost 700,000 downloads of the Windows installer. Plone 3.2.2 shows almost 200,00 downloads of the unified installer for *nix. If we take the data "as is," the graph looks like this with 3.2 dwarfing everything else.
I leave it as an exercise for the student to interpret all this. Does this mean lots of people download the Windows version to test the product before installing on a Unix box? Is Enfold Systems with it's toolset aimed squarely at Plone for Windows in the sweet spot for Plone consulting? And what is going on with downloads for 3.2.x? Your thoughts, please.
3 comments:
ppl on *nix use buildout.
I think the time between releases is a factor. I suspect that universal installer downloads peak after a release and then drop off, while windows installers have a more steady download rate. That should translate in a different universal/windows download rate depending on the time between releases. 3.3 will be a bit of an anomaly since it initially saw very little pickup; people appears to say on 3.2 for a long time.
A lot of people do indeed use buildout directly and not any of the installers. So you could aggregate the numbers from the different PyPi downloads (for example http://pypi.python.org/pypi/Plone/3.3.5) and add them as a "buildout" column.
Post a Comment