Redmine

My view point should be taken with a grain of salt since we're also affected by these practices from a business perspective, but I'd still like to share my point of view:

In general, depending on your plan level, we offer installation of arbitrary Redmine plugins on Planio. Unfortunately, this often leads to confusion and disappointment since clients take the screenshots and descriptions of the plugin directory on redmine.org for granted and believe that every plugin out there is of the same high quality/stability/state of maintainedness (is that a word?) as the core Redmine code base itself.

As we all know, this is - unfortunately - not always the case for plugins.

Thus, before we install a plugin at Planio we thoroughly vet it from a stability, code quality and security perspective. In many cases this leads to us having to fix/patch it in which case we contribute that back to the community (by the way, we have working forks of many - otherwise unmaintained, unstable or insecure - plugins; see Curated list of production-ready Redmine Plugins & feel free to use them). In a large number of cases it also leads to us declining a plugin altogether, be it for stability/security or (as is the case for EasyRedmine, e.g.) licensing reasons.

In those cases, we're often in the unfavourable position to have to explain clients that a plugin is not fit for production use, even though it is listed on redmine.org.

The addition of ratings to the plugin directory are already a great improvement over the previous plugin list!

From my point of view, I would really like some additional form of classification like commercial vs. free and also maybe some form of "authoritative review". For instance, it could be implemented as a score, or simply a yay/nay flag which only the Redmine maintainers/contributors (or whoever is knowledgeable and willing to) would be allowed to set. And yes, I would volunteer to help set these flags, or give these ratings, since we're doing it already if our clients request plugins.

So regarding the statements you requested, Jan:

• Remove all plugins that only work in some kind of fork

Agree, or at least mark them as such.

• Mark all commercial plugins with a very, very big "$" sign - or remove them from the index

Agree. I think marking is fine. In addition maybe we should tell people that over the top advertising with big red buttons and such as on http://www.redmine.org/plugins/redmine_luxury_buttons is not okay... In that case, I think they should buy Google AdWords which in turn at least benefits Redmine financially.

In addition I would propose to introduce a flag/icon/tag/whatever that says something like "Fit for production" with an explanation like "This plugin has been reviewed by a Redmine team member and seems fit for production use with regards to security/stability.", maybe with the addition of "No warranty, since it's open source, etc."

As I said, these are just my two cents. And my views are somewhat biased since we're not only a Redmine community member but are also taking part in a "commercial" way... If you disagree with my points, or have questions, please let me know.

Thank you Jan!

In addition I would propose to introduce a flag/icon/tag/whatever that says something like "Fit for production" with an explanation like "This plugin has been reviewed by a Redmine team member and seems fit for production use with regards to security/stability.", maybe with the addition of "No warranty, since it's open source, etc."

If plan.io does this anyway, then I'd say it's a good idea, otherwise there mean additional work for the already small team.

BTW: I didn't set this issue private on purpose, so that the discussion is transparent for all users...

Jan and Jan (sorry, couldn't avoid it), thank you for finally bringing this to the feature list :) As I started the (forum) thread, I just want to also post my two cents here.

If plan.io does this anyway, then I'd say it's a good idea, otherwise there mean additional work for the already small team.

This made me happy :) It's a great idea and could get even better if those "fit for production" gets featured somehow, like Google Chrome Extensions "staff pick". It's like saying "No strings attached, but they're good".

I have two more considerations, if applicable:

• Categorization: sometimes people looks just for wiki extensions, as an example, and it's hard to dig through 16 pages (or more) of plugins and descriptions to find out those who can be a wiki extension or not. Categorization or tagging can help a lot;
• Quality: as pointed out, plugin quality sometimes can be biased. One example I can point out myself was when migrating from Redmine 2.2.0 to 2.5.2 and the huge amount of effort I had to clean out the database from unused columns created by plugins in Redmine's core tables (I personally think that any plugin that modifies core tables is a bad plugin, but that's another point). So, plugin ratings and comments are available, but are these information used to something else besides visual user feedback? I mean, good feedbacks and a possible "add this as a (3rd party) feature" program could boost plugin authors to push harder on quality. As an example, I bring this up because Qt Framework uses something similar (providing Jira access, code repository, CI, etc) for developers who creates new features with their own efforts (and apart the main roadmap) and IMHO it works pretty well, without, of course, bloating things up.

My best regards,
Richard.

I'm not in favor of removing entirely paid plugins as long as it's made very explicit (same for a free plugin that depends on a paid one). That said, I'd really like we remove plugins that are incompatible with standard redmine, I think it just hurts redmine community as a whole.

That said I agree on nearly everything, and I especially like the idea of reviewing plugins depending on security/stability/production use criteria as proposed by Jan (which means I should polish my own plugins before listing them in the "plugins" section hehe)

Maybe JPL would agree if we share, even privately among contributors, the "Plugin directory" plugin and propose him some steps in this direction ?

Jean-Baptiste Barth wrote:

I'm not in favor of removing entirely paid plugins as long as it's made very explicit (same for a free plugin that depends on a paid one). That said, I'd really like we remove plugins that are incompatible with standard redmine, I think it just hurts redmine community as a whole.

Exactly.

Maybe JPL would agree if we share, even privately among contributors, the "Plugin directory" plugin and propose him some steps in this direction ?

Good idea, I'll propose that and discuss it with him once he's back.