Dev-starred items in the add-on server

[Jozrael]

Searching idly through the add-on server for something to while away an hour or two can be daunting: there is quite a lot of stuff on there that isn't on par with mainline, but quite a lot that is. It'd be nice if there was some sort of quality assurance that the devs could place on an add-on according to some criteria, whether it was dev vote, forum vote, or something else. It would just be nice to not have to hope that whatever we picked was finished/lacking severe spelling/grammatical errors, or other such flaws.

[Baufo]

(moved to "Ideas")

[Iris]

Searching idly through the add-on server for something to while away an hour or two can be daunting: there is quite a lot of stuff on there that isn't on par with mainline, but quite a lot that is. It'd be nice if there was some sort of quality assurance that the devs could place on an add-on according to some criteria, whether it was dev vote, forum vote, or something else. It would just be nice to not have to hope that whatever we picked was finished/lacking severe spelling/grammatical errors, or other such flaws.

The basics of this would be rather easy to implement IMO. But it could have some nasty consequences for the starred authors and non-starred authors. Think of what happens when you are working in a office, and some of your co-workers are more popular than you - many people find this frustrating when they have made such a great effort to do something good within their own limits. This could very well apply to add-ons if we had such a rating system. Some people excel at writing WML, but suck at storyline because English isn't their native language, they are affected by syndrome X or Y, etc. And the reverse.

Some people that get congratulated by an authority (in this case, we the so-called developers) also take their words too seriously, get lazy and their work starts deteriorating due to lack of proper maintenance/effort in making it stay in shape. This is of course nothing Wesnoth-specific. I have seen it happen (and has happened to me too) often in real life.

I'm just one developer, and of course my opinion in this matter does not represent, by any means, that of the whole Wesnoth project.

[Corvvs]

I think the downloads count already sort of serves the same purpose. Could be neat, though.

[Baufo]

I think the downloads count already sort of serves the same purpose. Could be neat, though.

No it does not, in no way. The addons on the top of the list always have the highest download count because people are to lazy to scroll down.

[Corvvs]

good point.

[Urs]

Hmm...

How about a rating feature, with, say five stars?

[Jozrael]

The question then is whether the devs or the users get to place the stars, or whether there is two separate ratings.

[Baufo]

How about a rating feature, with, say five stars?

The problem with such a rating system is that the majority of people are idiots.

[Urs]

The question then is whether the devs or the users get to place the stars, or whether there is two separate ratings.

The users and the devs, same ratings. AFAIK Devs are human too.

The problem with such a rating system is that the majority of people are idiots.

Well, yeah, but I like to think people who play Wesnoth are slightly smarter. Furthermore, it'd for the idiots, by the idiots.

[TheArchitect]

perhaps instead of a rating (such as 5 stars) we can add simple labels like "Completed" or "Developer Edited." They can show users which add-ons are'nt works in progress and guarantee relatively good spelling/grammar. Also, since they are based on fact, not opinion like a rating system, peoples' feelings won't be hurt.

[Ken_Oh]

I'd rather the big BWH concept of tabs/labels/sorting/etc. be done if anything is going to be done.

[Urs]

The thing is - five stars is simpler, easier to understand and more obvious. Tags, labels, and so forth aren't.

I do have the feeling this has been proposed before, though.

[TheArchitect]

five-stars is also innacurate because it is based on opinion. 75% of the wesnoth community might greatly dislike an add-on but that doesn't mean the other 25% will. However, because of it's rating, lots of people won't even give it a chance. Also, I don't think wesnoth should drive off new contributers with bad ratings just because their very first contributions left a little to be desired.

And honestly, do you really think it takes a genius to figure out what a label like "complete" or "incomplete" could possibly mean when applied to an add-on? Sure, five-stars does give a little more information about which add-ons are more popular, but it has serious drawbacks. Tags and labels on the other hand do not.

[Viliam]

I think the downloads count already sort of serves the same purpose.

When I download the add-on and I like it, it increases the download count exactly the same as when I download the add-on and don't like it. But yes, if I like it, I will probably download it again after installing a newer version of Wesnoth, and I will also recomend it to other people. So the download count approximately corresponds to popularity, but a bad add-on with very cool name and icon could have this number higher than its popularity.

five-stars is also innacurate because it is based on opinion. 75% of the wesnoth community might greatly dislike an add-on but that doesn't mean the other 25% will.

If 25% strongly like something, and 75% strongly hate it, it would have the same average result as if 100% think it is poor but somehow acceptable. And this is not what users need. In the first case, if 25% strongly like something, it should receive high rating and a flag, because those 25% want to find it, and the others want to avoid it. In the second case, if nobody likes it, then nobody wants to find it, so the rating should be low. The goal of rating is to help people find the content they will probably like.

This rating system is similar to literature rating, for example. If someone says that "Asimov's Foundation is a great book", it means that it is very popular in some minority (sci-fi readers), despite the fact that average person (not a sci-fi reader) probably would not like it. If we have a tag "sci-fi" and a rating "great book", we understand what it means, and a person who does not like sci-fi will not buy it, but will not say that the rating is wrong.

I think if we are going to measure the quality (not the % of players), five stars are probably too much. How many degrees of quality of Wesnoth add-ons are you able to judge consistently? I would say four (0 to 3) are enough: 0 = "it's really bad (unfinished, full of bugs)"; 1 = "it works, but it did not impress me"; 2 = "not the best, but nice"; 3 = "very good, I would recommend this". The best rating given by a significantly large majority (10% of players?) would be the result rating.