Garden Plugins Site - User Ratings

Stash

I'm just thinking ahead a little to Garden and how I think the generally excellent addons site could be improved for it. I would love to see some kind of user rating on the plugins, ideally showing both for the current version (which resets each time a new version is released, maybe with a history?) and a long running rating for the entire life of the plugin. I think this would help people find decent plugins and also let them know if the current version is any good - would this be a lot of work to implement on the addons site? What do other people think?

x00

I think it is maybe part of the solution. Also some of it is down to good feature and a culture of conventions in garden (if you want greatness in a free environment then conventions can be invaluable). The main problem is not that plug-in could be terrible, because that would always be the case but it is the fact the plugging can so easily conflict with one another no mater how good they are. I am going to discuss this in due course. Things like namespacing your JavaScript, and some smart scoping ideas php without having to rely on namespacing which is not the best idea thinking it through. These shouldn’t come at the expanse of the coding freedom. I would probably use the MVC paradigm form my major plug-ins. But really there is no reason why you absolutely have to. It could be a one script job, nothing wrong with that. However next up was going to address those damn css titbits next. [-Stash-] you destroyed my schedule :p.

TomTester

Second the motion (stashes) - too many questions on add-ons pages that say "does anyone know if ti still works in version x.x.x?" Re plugin quality: might I suggest a well-commented bare-bones demo plugin download which includes a fixed structure as a 'convention', or perhaps even something better... PLUGIN QUALITY CONTROL PROCESS - The plugin quality control process would involve an experienced team of reviewers to check new plugins. - Plugin authors could submit their plugin for certification. - Approved plugins would be certified (this garden plugin is 'weed-free' and compatible!) - Certified plugins could be sold for a nominal fee ($3 per plugin? 30 day trial?) - Plugin certification would either cost $$$ (paid by author, author would retain X% of revenues) or alternatively be 'free' but plugin control team (or garden/vanilla/mark) would share in revenues. Something like that. You'd be the first forum to have this, I think. I'd be more than happy to pay $60 for 20 plugins, knowing that they were tested to work together. Hell, I'd pay $100... (still less than my hourly rate)

x00

TomTester I really don't like the idea of charging for my plugins, and trial periods and scripting don't work. I think it is ok to have a donations button, and also pay someone to do a plugin that hasn't been done, that others can enjoy after. otherwise it become one of those crass convoluted "frameworks" like joomla and the would be the death of Garden. The quality will go down not up. Like I said the biggest problem is plugin conflicts. You can't expect the developer to test every combination of plugins. There is more to be done on the front. The conventions go hand in hand with documentation. I also think it would be a really good idea to have a plugins competition on the launch of Garden and Vanilla 2. have different categories, themes, extensions, etc. to be voted on by the users. With different accolades, such as creativity and originality, most useful, etc.

x00

also sometimes the bad one have some good ideas behind them

TomTester

@x00: I like the idea of a plugins competition. Let's not limit it to plugins built, but also include a vote for plugins needed. It's funny that you say "You can't expect the developer to test every combination of plugins." That's EXACTLY the problem. You can't expect him/her to. You might convince them too with the right motivation (e.g. money) or you might offer them help and do it for them (i.e. certification/support). You can attract/retain better talent if things generate money for them (even if it's just a little), be it a plugin author a plugin reviewer/tester, or Mark himself... Just to be clear: - I'm saying *I'd* PAY for plugins that HAVE been tested (you don't have to). - I'm saying *I'd* prefer paying a little over spending my own (expensive) time debugging. - I'm saying *I'd* prefer well-tested plugins (you might not) - I'm NOT proposing that ALL should be paid, - I'm NOT proposing that ALL should certified - I DO expect that a 'certified plugin' program would benefit the overall quality of the plugins and stability of Vanilla/Garden It's a fallacy to believe that you can do all this stuff for free. This is stuff I believed as a student, when money was no object and my time was free, but in real life that simply is not the case. This is also obvious on these boards. Just go back a couple of years and you'll find that very talented people, once very active, have "moved on" (gotten babies, new jobs, etc.). For some of them, a little money might have been enough of an impetus to stick around longer or maintain their code for the benefit of their own wallets (and the community). Maybe what I outlined was a little too complex, but judging by the amount of time I recently spent to get my new site up and running, the large number of "small" changes I had to make to various plugins to ensure they worked together, worked with Vanilla 1.1.5a, etc. etc. made me realize more than ever that a change is needed. TT PS Agree with scripting/trials being annoying... but there has to be some impetus to pay for the work of others.

x00

I am not going to argue with you about the merits of paid versus unpaid. All I can say it doesn't mean the quality will go up. Also it is unreasonable to test against every other combination of plugin (and all their version), because it is a run away train. it is not going to happen. Like I said smart design in garden and convention can addresses common issues. Also just because a plugin has got bug, or goes through beta testing doesn’t make it bad, that is part of the process. There is a fine line between having standards and alienating your contributors.

Stash

I just like the laid back approach of allowing people to give a little "one click2 feedback on the current release of an extension as I think that will provide some immediate useful feedback. If you wanted to make it slightly more involved on the site backend you could track who's downloaded what (and which version) and ask them to give feedback on them - kinda like ebay does with purchases/sales. Just a thought. WRT the money suggestion - I'd personally rather keep it free here - you can always sell your own extensions elsewhere if you want... also, consider that if you really want an author to continue with an extension there's nothing stopping you from paying them whatever you feel is fair enough to motivate them.

scubaguy

the Joomla extensions site does a nice job with this (example: http://extensions.joomla.org/extensions/communication/forum/7256/details)

Stash

Yes it does (although not in a very pretty way :D), but I can't see if that's "per version" voting or just a general "all versions" feature. I think a per version combined with an all versions is important.

x00

Please don't make Garden like Joomla, not as a framework or an addons repository. It really will be the death of it. I won't even go into how much of a convoluted framework joomla is, however their extensions are a good example of how the quality is not terribly good, in fact it is often the ones that try to take money off you that are the worst. The reason is that thought of bit a cash attract people who really haven't go the first clue how to program. One thing that is definitely true is how you present yourself creates the culture that surrounds. Joomla makes an awful lot of fuss about being open source, as if they are doing anything out of the ordinary, but in reality their culture they create simply attract fly by night botchers, who just want to take cash of you. There is an awful lot of paid extensions, that wouldn’t be paid on another framework. That will automatically turn clued up people away from the framework. There is actually less accountability on Joomla than there is here, despite ratings, etc.

TomTester

Let me restate this one more time: I suggested for-pay plugins to raise cash for quality control (testing takes time, time == money) by 'vanilla experts' and a virtual 'seal of approval' for certified plugins that you could depend on when making a decision to purchase. If it works for the Apple iPhone store, then why not here? (let's skip the odd anecdotal exceptions like baby-shaker for now). User ratings can work in lieu of an official 'seal of approval'. It just moves the task of validating a plugin to the "crowd" vs a small set of experts, which adds a layer of complexity (or stupidity). As Amazon shows, some people give bad ratings when things don't do what they expect or need, when they are not smart enough to read or follow the installation instructions, or the box the widget came in was the wrong color (amazon only, no boxes here) etc. etc. Hence, whatever rating system you implement, let's make sure it has a ratings moderator feature too and/or some sort of rating weights attached (e.g. a function of length of membership, number of posts, add-ons published, etc.) Because if [-stash-] gives an add-on 5 stars, I would not want the vote of a zero-day noob who rates it a 1 star because the CSS was the wrong color for his theme to lead to an "average" of 3 stars.

mumpitz

Hmm...

Is a voting system really necessary? When I look at the extensions, which are available right now, I don't see any two with the same functionality (except the rss-feeder and 1 or 2 others). If I want a certain functionality I can only chose one addon. What use is it then I know it has 1 (or 5) stars?
Even if lots of addons come up, I don't really see the use for a rating system. When I need a certain functionality I look through the addons, read the description of the author. This gives me an idea if the addon will satisfy my needs. Not some rating.

The only reasonable rating I can think of: "It works!" <-> ("It kinda works") <-> "Its not working". This should, of course, be coupled with the current version of Vanilla.

In my opinion an elaborate system to advertise similar addons is of much more use. For instance, when I am viewing the page of the FeedThis-Addon it shall give me the hint that CrudeRSS and FeedPublisher might be worth a look since they might meet you needs even better. Thus making it easier to find similar extensions.

SubJunk

mumpitz one of the advantages in that scenario is that while now there might be just one extension claiming to offer a functionality, it might break the forum. If it has 1 star on the website then people would know it doesn't work well so will be able to avoid screwing up their forum.

TomTester

Not to complicate things, but the "doesn't work" verdict would of course very much depend on other parameters, incl. but not limited to technical expertise of the user, technical setup, installed add-ons, etc. etc. I recently read an article (can't find it) that one of the smartest things M$ has ever done is adding Windows Error Reporting. Perhaps a similar "tool" or active "reporting option" could be added to Vanilla2 (or older versions) to assist in improving the quality. One possible implementation would submit (upon request or automatically) critical install info to Vanilla/Lussumo, e.g.: version of PHP, MySQL, installed modules (partial or full PHPInfo), site address, errors thrown, list of bling and bling versions installed on the forum (add-ons, theme), list of bling presently active on the forum (some installs add things to the DB, but deactivating doesn't remove those additions), hosting company, CRC of all files (to detect mods), etc. etc. Over the long run this could be used to see if there are "patterns" that might point to problems. In the short run, this info could be provided to friendly forum participants to assist in resolving the problems of a particular user. http://en.wikipedia.org/wiki/Windows_Error_Reporting

mumpitz

mumpitz one of the advantages in that scenario is that while now there might be just one extension claiming to offer a functionality, it might break the forum. If it has 1 star on the website then people would know it doesn't work well so will be able to avoid screwing up their forum.

Yeah... thats what I meant with the only reasonable rating I can think of. However, then the rating should be done for the question "How is this extension working with your installation?" ("Does it break your Vanilla?" sounds so negative ) and not "How do you like this extension?".
In my opinion, somebody how wants to use an addon should examine the description thoroughly and scan the comments of that addon. And if you do that a rating is just of very very little use.

Anyways... while reading TomTesters post another idea which will probably resolve the issue of "is it breaking Vanilla" came to my mind. The checking for upgrades script of Vanilla could send the lussomo upgrade service which version of Vanilla is used and which addons are activated. With those informations it should be pretty easy and reliable to signal if an addon works or not. This will also be a more reliable source for the popularity of an addon compared to a rating.

Stash

Hmm, I quite like that idea of showing "Installed ##" and "Enabled ##" statistic, that would be great :)

TomTester

@mumpitz: love the idea! simple and effective...