[interface] New Filtrer to find Add ons (New idea!)

[doofus-01]

Rating systems work. They have worked for many games with add-on content.

Could you provide examples? A google search for "rating system for add-ons" quickly turned up this thread, a similar one for mozilla, and some other mostly irrelevant stuff.

EDIT: Oh, answered in a parallel thread. Too many of these things...

[lokiju]

Rating systems work. They have worked for many games with add-on content.

Could you provide examples? A google search for "rating system for add-ons" quickly turned up this thread, a similar one for mozilla, and some other mostly irrelevant stuff.

EDIT: Oh, answered in a parallel thread. Too many of these things...

Sorry for the confusion!

[SkyOne]

Ooops. It is a part my fault. The topic is from: http://forums.wesnoth.org/viewtopic.php?f=6&t=33139
Excuse me, doofus-01.

lokiju, would you like to discuss it here?
If so, you ought to request to lock the thread on User's Forum with the link of this thread.

[IPS]

Also I know now another form how the people can rate the things: by the simple count of positive and negative votes; incluiding a stuff like adding "favorite" option.

There is a new idea now, an example for this is if were a form to the people could vote for those add ons what really do like or dislike to them. Also there can be an special botton like "following" or "favorite". If you dont know the add on or you dont like or dislike an add on, a simple stuff what the player can do is ignoring the votation for the add on.

With this idea, the followers of the add on can find themselves by the only add on server and then enjoy they alone for their favorite add on and not rejecting the game because its so very impopular (liked to me with Breaking Ground add on, I like it a lot, but nobody want to play it )

[Atz]

Ugh, not in favour of ratings. I've seen other communities where ratings of user-made content were a constant source of drama. Some people whine endlessly that their work isn't universally loved by everyone; others go around rating things down because the creator beat them at a game, or argued with them on the forum, or has a higher rating that their own work, or just because they like annoying random people; and yet others were constantly trying to create fifty accounts to vote up their own work, or getting all their friends to sign up and say how awesome their stuff was without even trying it. Then there are the people who spam the forums and IRC trying to get their work rated, the people who constantly make threads about how it's a travesty that [UMC1] is the highest-rated when [UMC2] is clearly better, the people who complain that some content gets high ratings purely because its creator is popular on the forums, the people who get into huge arguments because one of them thinks 8/10 means "great" while another thinks it's "mediocre", and so on, and on, and on.

The other problem with Wesnoth in particular is that many add-ons are under active development, and complaints about bugs or other problems can rapidly become misleading because said bugs have been fixed.

[IPS]

I'm right with your point. Also the people tend to be very vengative for negative votes... and that also can make a chaos when are talking about the add ons.

But it must not be bad if we eliminate the negative votes, making it for only positives and favorite votes, no vengance is posible, but also the multi counts for voting and the inactive add ons stuff may not be fixed with it.

But of course I think there can be a correction for atleast the most important problems what may it cause, but we need more than a person (opposition and followers of this idea) to see how good can be this idea. Let's wait nvm.

[Reepurr]

There is a new idea now, an example for this is if were a form to the people could vote for those add ons what really do like or dislike to them. Also there can be an special botton like "following" or "favorite". If you dont know the add on or you dont like or dislike an add on, a simple stuff what the player can do is ignoring the votation for the add on.

With this idea, the followers of the add on can find themselves by the only add on server and then enjoy they alone for their favorite add on and not rejecting the game because its so very impopular (liked to me with Breaking Ground add on, I like it a lot, but nobody want to play it )

I quote the FPI.

There should be some sort of rating system for the add-on server. (Reason: Proposed multiple times, it has been rejected because a subjective high rating does not make it a good add-on.)

The Devs apparently don't want any rating system for add-ons. Technically, this can be classified as a rating system, because the more people "favourite" most of the add-on server, the more an un-favourite-ed add-on is classified as rubbish. Also, as an add-on gets older, more people will favourite it, because more people will know it exists. A new add-on would not be marked as anyone's favourite, because nobody's played it. Which could mean that, despite the add-on being extremely good and, in fact, about to go into Mainline, nobody would play it.

Seriously, even if you don't allow people to rate an add-on bad, someone will just tally up how many people voted [Add-on 1] good and how many people voted [Add-on 2] good. And should the former have more good votes than the latter, then [Add-on 2] will look like it's effectively been voted 'bad'.

EDIT: I see that the FPI has been quoted before in this same thread...

[Neuromancer]

Unfortunately, some UMC developers are not amongst the brightest bunch and can’t follow standard software development practices.

For those who can read English and visit the forums, there’s this thread, but then we go back to the “not amongst the brightest bunch” part.

Use the standard version numbering scheme: 0.x or 0.x.y for incomplete versions and 1.x or 1.x.y only for versions which are complete (fully playable from start to finish without major problems). You can still attach other notes or labels of your choice, so versions such as “0.9.1 beta” or “1.1 (updated)” are fine.

Solution: Show the message saying this ingame the same way GPL announcement pops up.

Furthermore, I think that if the popularity of Wesnoth keeps growing and the number of add-ons keeps growing, eventually it will reach a point where we really need such a ranking system but currently I don't think it's needed.

I think this point was already reached. lokiju nicely stressed the importance of the rating system.

At the risk of saying something heretical, I'm going to disagree with the basic premise of the FPI #32. I think there are plenty of real world examples that show rating systems can be effective, even accounting for differing tastes. I think intelligent reviewers are capable of recognizing those differences and rating content accordingly.

Agreed.

There are not enough qualified and dedicated content reviewers, not enough for me to want such a system in place. If you don't believe me, take a look at WSRSW site and how few people bother to rank the content there.

Because nearly no one knows about that site, has crappy interface, and you have to register in order to rate.

The current state of the addon server is appalling. There is this endless loop of have-the-most-downloads <-> gets-the-most-downloads.
I think the basic premise we must follow when designing the rating system is this: Every rating system is subjective.
I think we have to avoid systems based on popular voting (see how it works in RL, populistic parties winning the elections) and systems based on selected voters, this will create elitist circle.

Proposed solution*:
Remove the "Downloads" sort option.
Make the initial listing of the addons completely random. -> Now we have fair ingame environment for the addons.
Make clickable urls ingame, encouraging addon authors to add links to their respective addon topics in addon description.

Next, promote users to create personal wiki pages in which there will be their reviews of the addons. Make simple rule: If content version is lower than 1, users are not allowed to rate that content on their wikis. In this way unfinished projects will be not hampered by initial negative reviews. Next, there will be two wiki pages: one containing the list/links of reviewers (nicks color-coded like on forums), the second list/links of the reviewed addons, again randomly sorted, with search function.

I also propose to remove Addon feedback section in forum. Project feedback is usually being posted in development topics.

*I think the good solution doesnt exist because of the subjectivity thing, there are only more or less bad solutions.

Edit: typos

[doofus-01]

I also propose to remove Addon feedback section in forum. Project feedback is usually being posted in development topics.

I don't think it hurts anything, though scattering posts between two threads per add-on isn't so great. If the threads were strictly feedback with the template questions, it could be useful, but the distinction between development thread and feedback thread is too subtle if you don't look at the forums regularly, so the add-on threads are all muddled.

[Gambit]

I think a big problem there is that add-on developers don't know how to manage their threads. You need to keep the first post up-to-date as a sort of index page.

[doofus-01]

I can't speak for authors in general, but I know I had posted links in my first feedback post to the development thread and asked for bug reports to go there. To no avail. How would posting and updating the version numbers change that?

We can't really manage threads, we can only post suggestions which will go right over the head of anyone who doesn't browse the forum enough.

[TheCripple]

The original request was for a way to filter add-ons based on an objective measure of popularity. Something like Downloads/Time. This wouldn't fall into the FPI of having rated add-ons, because it regularly mixes things up. New add-ons would be able to spike up in popularity if they get 200 downloads their first day. Older add-ons would sink back down when they stop getting downloaded at such a massive rate.

I'd recommend against downloads/time, it inherently favors newer projects. Essentially., what I'm stating is essentially that time grows faster than downloads, leading to a decrease in Downloads/Time. This happens for a few reasons, and its probably best to look at each variable on its own.

Time. This increases at a constant rate.

Downloads. Modeling downloads involves looking at a few things. Essentially, you have the amount of people who would download the mod without the Filter affecting it, relative to the size of the community as a whole. This stays relatively constant. Then you have the amount of people within that group who have already downloaded it. As the amount of people in the group who have downloaded it increases, the amount of people who would download the mod relative to the size of the community gradually shrinks, though the influx of new people reduces the rate at which it shrinks due to the growing active community. However, the rate at which something is downloaded is linked to the amount of people in the group who would download it who haven't yet, which is an inherently shrinking rate.

Basically, we have a (rate of download)/(rate of time) decrease with time, which leads to Downloads/Time decreasing. This essentially means that Downloads/Time favors newer material, which I view as a problem. However, that it is a problem is very much subjective, we may want this to favor newer material, and if so that is a happy side effect.