This is just to let you know that we have reviewed our plugins downloads tracking code and data. It was confirmed that there were some little bugs well hidden in the corner that caused suspiciously high and false downloads figures being reported for our plugins. These bugs were fixed at the end of September. Unfortunately we were not able to propagate the fix on existing aggregated data so we had to effectively reset the download counters. Starting from October 1st, 2015, the downloads figures should be significantly more reliable.
As a consequence, we have slightly modified the Plugins directory UI. Where "total downloads" were reported previously, we now display sum of "recent downloads" from the last 90 days. So these stats now represent the current trends rather that overall downloads activity.
It turned out this can be seen as actual improvement as it allows new popular plugins to stay side by side with old and well established ones on the stats charts.
Additionally, the chart of plugin version downloads by month was improved to better visualise the history and trends of interest for particular plugin version. And, that chart's legend should now be a bit more compact for plugins with many versions available.
Hope you enjoy these news stats and thanks for understanding.
It seems to me that the most useful as well as the most reliable metric for plugins is the number of sites on which they are currently installed, but this metric is not used in comparative data, while both the number of and downloads (Category stats) and favourites (Plugin listings) are.
The number of downloads statistic is very sensitive to how often and how recently the plugin was updated, and somewhat penalizes more reliable and stable plugins versus others that are frequently updated (perhaps to fix bugs).
The favourites statistic is a bit of a popularity contest, and can be heavily influenced by how active developers are in promoting their own plugins (so I've found anyway ).
From the perspective of an administrator deciding what plugins to install, it would seem more useful to list plugins within a category by order of installed sites, rather than how many people have marked them as favourites (e.g. Atto plugins list).
Alternatively, the category statistics (e.g. Atto editor plugins) could include a graph showing plugins ordered by number of installations either instead of or as well as by number of recent downloads.
I think this data would better show the most useful plugins, as indicated by being currently installed, preferible by Moodle branch.
This would help Moodle admins who are thinking about (contemplating) installing additional plugins, It would also help to direct Moodle documentation writers and Moodle translators to know which plugins are most used in the real world.
Thanks in advance.
That's a very good suggestion - thanks Tim. I will make that happen.
Meanwhile, I was able to dig out the following stats for your qtype_stack plugin:
plugin qtype_stack: 540 - plugin qtype_stack on moodle 3.1: 1 - plugin qtype_stack on moodle 3.0: 89 - plugin qtype_stack on moodle 2.9: 190 - plugin qtype_stack on moodle 2.8: 82 - plugin qtype_stack on moodle 2.7: 115 - plugin qtype_stack on moodle 2.6: 44 - plugin qtype_stack on moodle 2.5: 10 - plugin qtype_stack on moodle 2.4: 3 - plugin qtype_stack on moodle 2.3: 6
Based on data from January 2016. Hope this helps.
Thank you David. That is really helpful (even if it tells me that it is important to still support STACK on Moodle 2.7, so I actually need to fix the bug that only happens with that version ) .
It would definitely be useful to have this information in the plugins DB, so I have voted for MDLSITE-4399.
The main Plugins page seems to indicate that there are 1426 plugins (adding all numbers inside parentheses).
However, there is also a piece of text that says "1151 plugins" on the top right of the page.
The overview page states that there are now 6008 plugins listed in the Favourite plugins report, and it has 121 pages with 50 plugins per page, indicating 6000+ plugins, but pages 20 to 121 are empty, so there are only 950 listed. Are the remaining plugins those with 0 favourites?
I think I can answer your questions, through a brief bit of investigation:
- The discrepancy between 1426 and 1151 comes about because you added all of the numbers in parentheses - thereby double-counting any which are in subcategories. The number in the heading at the top of each block (Activities, Availability conditions, Blocks, etc) includes any which are in subcategories - but not all plugins in each category fall into a subcategory (some sit directly within the top-level category, i.e. Activities which stand alone rather than plugging in to one of the core Activities).
- The 6008 appears to be the total number of "favourites", rather than the total number of "favourited plugins" - that seems like a bug to me (can be logged in the tracker under the MDLSITE project - a quick search doesn't seem to reveal any existing reports of this issue). The ~100 extra (empty) pages will be because the incorrect total of 6008 is being fed into the table library, so it thinks that there are 6008 records to output rather than the actual number of plugins which have been favourited.
Just to clarify my earlier post:
- There appear to be 1151 plugins at present (the numbers for the top-level categories sum to this figure, which matches the top right corner of the page)
- The figure of 6008 appears to be the number of reviews rather than reviewed plugins; this incorrect figure is being fed into the table library, resulting in ~100 empty pages.
I've also gone ahead and reported it to the tracker: MDLSITE-4400