List of Plugins sorted by number of install

Hi,

This is a feature request.

It would be nice if we have a page that list all the plugins sorted by the number of installs. I’ve checked plugin - Discourse Meta but I couldn’t find a way to see which are the most installed.

I want to launch my fist site and I would like to know what plugins are the most needed, and I guess all the newbies like me need to do the same.

Drupal modules page sort all modules by the number of installs: https://www.drupal.org/project/project_module

Ohh no!

  • Discourse doesn’t send statistics about plugin installs to a master server?
  • Most Discourse plugins are hosted on Github, and maybe there is no way to retrieve the number of downloads (a proxy for the number of installs)

Why would this be useful in practical terms? Just because others do it? You are the first to ever ask about it, to my knowledge.

3 Likes

Sorting by popularity is not uncommon for any type of extension. A number of active installs usually informs about reliability. A broken extension would go down pretty quickly.
I wonder if tracking active installs is even possible though.

1 Like

I have the belief that decisions are best when they are based on whether or not something is the best tool for the job. eg. does it do what it is wanted it to do, without needing hacks and addons and without unnecessary bloat?

I suppose popularity could give some indication of its fitness. As could how actively it is being updated. But neither are perfect. Neither popular nor recently updated equate to best choice.

I think some type of rating feature could work, but even that could be gamed and misleading.

2 Likes

It is not currently possible, there was some brief discussion about this in this topic:

I agree that having some install statistics would be good though - as well as helping users decide whether to install a plugin, it could help developers target their time based on their most used plugins. Unfortunately implementing that from a political/privacy perspective is not a simple matter.

4 Likes

I suppose it would be useful if we had (once we have?) a lot of competing plugins i.e. plugins doing about the same thing, as is the situation with wordpress plugins. There, the number of installs often saves you hours of research (or limits that research to a few short-listed plugins).

In the case of discourse plugins, the challenge is probably more to figure out whether a plugin is used by a niche audience and possibly no longer maintained or widely used, making it a relatively safe thing to use.

2 Likes

And my position is that outside of extremely rare exceptions, only plugins we officially build and support should even be considered. Which makes the rest of the stats kind of irrelevant at this time.

But i suppose you’d want to encourage a vibrant community of plugin developers to add features for all occasions.

But balanced with the need to control quality so that bad plugins don’t drag discourse down with them.

In this conflicting goal, some form of statistics will be a great solution.

4 Likes

I find the most interesting statistic not to be usage, but rather “being around to take care of plugin”

  • When was the last commit?
  • Is plugin owner active on meta?
  • Is plugin owner responding on plugin topic?

I also think this is definitely homework everyone should do prior to taking on a 3rd party plugin

Last thing you want to do today is commit to a plugin nobody will be maintaining

9 Likes

I feel the importance of this bullet really depends on the plugin.

Take something large like https://meta.discourse.org/t/topic-list-previews/41630. It’s a large plugin and touches lots of areas of core, thus requiring frequent commits. Compare that with GitHub - discourse/discourse-microsoft-auth. Its last commit was 8 months ago and it works just fine (in fact it’s installed on our business tier).

6 Likes

I think there are a number of pitfalls here, most of which have already been mentioned.

There’s a plugin for that :sweat_smile:

I agree.

I think there’s a distinction to be drawn here between the completeness of the feature set of Discourse vs a CMS like Wordpress or Drupal. For most people standard Discourse, perhaps with a theme, is all they want and / or need. When you install Wordpress or Drupal you typically install a handful of plugins just to launch the thing.

Having a rating system or stats creates a new social ‘norm’. You may not be considering a plugin at all, but if you see that others have given one 5 stars or that everyone else is downloading it, then you wonder what you’re missing out on.

Speaking for myself, it’s not clear to me that it would be a good idea to encourage this kind of norm creation. At least at this stage.

And as @sam pointed out, perhaps the most important ‘stat’, i.e. is the thing actively maintained, is hard to quantify beyond a label.

7 Likes

If it is official and published by us the answer to that is always YES, 100% of the time.

3 Likes