Download count being an implicit ranking system

[Insinuator]

A review system would be nice to have but it is *not* solving the problem.
You can't sort a list by data mined from the content of a review system easily.
And we must catch the first time user of the addon server.
Let's make sure he most likely downloads something that looks at least like it is worth to spend more time using
the addon manager.

We already can data mine to an extent. As the very title of this thread states, the current download count is a sort of ranking system, whether we like it or not. We don't have to get rid of that. Think of Reviews as supplementing, not supplanting. We put a Reviewing option into each add-on but leave the sorting system intact. If people only want to look at download count and don't care to read the reviews, ok. But if they are searching for a better system of determination, they have it right at their fingertips!

Oh, so let's not fire at the monster unless we can kill it one shot? Those would still have more use than throwing around random ideas in a forum thread

Here, here! Lipk is quite right. I would like to make a Feature Request if there is sufficient support. That, at least, is a small step in the direction of implementation.

[Adamant14]

Oh, so let's not fire at the monster unless we can kill it one shot? Those would still have more use than throwing around random ideas in a forum thread

Here, here! Lipk is quite right. I would like to make a Feature Request if there is sufficient support. That, at least, is a small step in the direction of implementation.

You receive my full support.

[Pentarctagon]

I don't think Tips is a good place for a link to the UMC Guide. If we want people to actually use it, the link should be located somewhere it would be more consistently visible.

Also, if it can't be clickable, it should be able to be copy/pasted. Short link or not, having to manually type out a url is pretty annoying imo.

[Crow_T]

Also, if it can't be clickable, it should be able to be copy/pasted. Short link or not, having to manually type out a url is pretty annoying imo.

Indeed. I can't help but feel that it would be a great idea to rewrite UI elements in HTML/CSS, allowing for easier expansion for mainline things by the community and UMC devs to make cool custom UIs.

[fabi]

A review system would be nice to have but it is *not* solving the problem.
You can't sort a list by data mined from the content of a review system easily.
And we must catch the first time user of the addon server.
Let's make sure he most likely downloads something that looks at least like it is worth to spend more time using
the addon manager.

We already can data mine to an extent.

What are you talking about?

As the very title of this thread states, the current download count is a sort of ranking system, whether we like it or not. We don't have to get rid of that.

I disagree, we have to get rid of the fact that high download counts don't match with high quality.

Think of Reviews as supplementing, not supplanting.

I already do. Reviews could supplement a rating system, but not supplant it.
And they alone would not solve the problem.

We put a Reviewing option into each add-on but leave the sorting system intact.

And that would solve which problem?

If people only want to look at download count and don't care to read the reviews, ok.

If I understood him right, this is a good example for what he meant with

Wesnoth has many strong points but team and users management are certainly not in them.

But if they are searching for a better system of determination, they have it right at their fingertips!

You must be kidding.

Let's see:

Our new player discovers Wesnoth, she likes the game and plays through some or most of the mainline campaigns.
The rest of them does not suit her, she decides to use the addon server to search for new content.

The first download is not a campaign, maybe the second isn't either.
Thus the new content is not appearing where she expect it.

She assumes the game to be broken (it is for free and some hobby people are driving it, right?) and waits for the next version with
a working addon system and more mainline campaigns.

Or she might ask here in the forum ( That happens from time to time, estimated number of unreported cases might be very high),
first problem solved.

Now she knows about the different types of addons and will choice the right one, "campaign" in her case.
Having to choice between 200+ different campaigns on the server she looks for a sign of quality or just downloads randomly or when she likes the description/icon/name/whatever.

Most likely the first few downloaded addons are again not suited for her.
She comes to the conclusion that all good campaigns are already mainlined and waits for the next release containing more campaigns.
Or she might discover your review system and spend a lot of time reading through reviews, most oft the time not understanding what the review tries to tell her since she is not yet familiar with Wesnoth's vocabulary.

Now, the problem explodes. Instead of viewing through a list of addons she has to view through a similar huge list of review lists.
Again with no hint which addon it is worth to read the reviews for. (Assuming she can read and understand English well enough)

Thus after a lot of time spend with the issue she might end up with a campaign she enjoys playing.

Does anybody want to suggest that this is a good way to handle users?

[Dugi]

fabi, you are perfectly right. I can't tell that you persuaded me because I thought so also before, but I strongly agree with you.

The way I see optimal at the moment is to use a combination of tags, ranking and reviews. Tags to possibly sort the add-ons to get those one is interested in (for example undead campaigns), ranking to get quickly to the popular ones (of the intended kind, if the tags are selected), and then to read the few reviews and try to get an idea what is the best. Even if the user doesn't understand much of the tags and reviews, she will still be pointed to the most popular ones to try a few and learn to understand the reviews and tags.

[H-Hour]

Does anybody want to suggest that this is a good way to handle users?

It sounds to me like what you're talking about is as much a UX problem as it is a ranking problem. One thing you need is some kind of UI intervention which puts a new user -- who doesn't know anything about Wesnoth's add-on ecology -- on the path to what they're looking for.

Currently, the game just loads all of the add-ons automatically. What if, instead of jumping straight to the full list, the main menu Add-on button opened a dialogue in which the user was asked to select the add-on types they wanted. This would implicitly inform the user that not all add-ons are campaigns by forcing them to choose the add-ons they want.

A quick mock-up of what it could look like is attached. I've just taken the categories from the "Options" menu and laid them out with short descriptions (obviously better descriptions could be made).

[Adamant14]

I really like H-Hours suggestion.
It would solve one of the issues which are handled here.

I can remember my first add-on downloads.
After playing through all mainline campaigns I wanted more, I wanted to play a UMC campaign.
One of the add-ons.
But nobody told me that not all add-ons are campaigns, so my first few downloads wasn't campaigns.
I remember how disappointed I was the, because I couldn't find the new downloaded campaigns nowhere.
I remember that I was about to quit.

H-Hours idea can prevent that.
And I guess it can be done easily.

[Iris]

I just committed a series of changes so that starting with the next development version (1.11.5) clients won’t bump the downloads count when upgrading or reinstalling add-ons that are currently installed.

[Dugi]

Good job. Is it done so that all the older versions will bump it (sending the information that it is just an update) or so that only the new versions will increase it (sending the information that it is a new download, that might be better, because otherwise the download count of the add-ons with a high number of downloads will keep affecting it until the 1.13 version)?
A reset of the downloads count might help too.

[Iris]

Good job. Is it done so that all the older versions will bump it (sending the information that it is just an update) [...]

Yes.

A reset of the downloads count might help too.

Nah, that can wait for the first 1.12 beta at some point in the future.

[Dugi]

That means that the current problem will remain till 1.12 beta. Better than nothing, but still...

[Groggy_Dice]

It seems like the sentiment in this thread leans towards going back to the practice of wiping the dev server at the beta stage, as well as deleting broken add-ons. I'm not for this. Although the thought may be that those who don't stick around until the dev branch hits beta aren't serious, there is real lost content.

For instance, Guardian Order still has a version for 1.4 that notes that there are "2 scenarios completed thus far." However, if you look at the archive.org waybacks for 1.7, you can see that the attempted port had "4 scenarios completed thus far." But those two new scenarios are apparently casualties of the server wipe.

On the other hand, the situation with 1.9, where there was no new instance started for the stable series, and broken add-ons like Love to Death from early in the 1.9 development not only linger, but get lots of downloads because they're first on the list, is not ideal.

Personally, I'd like to see something like what happened with 1.5/1.6. A new server is started for 1.12, but 1.11 remains up. Add-ons that get uploaded to 1.12 are removed from 1.11, while those that don't get updated remain. If someone uploads a broken add-on to the stable server, they are booted down to the old dev server, but not deleted entirely. This way, they are removed from the choices that visitors are offered, but are still available for those who really want them.

Of course, it's not my storage or bandwith or admin time.

[DranKof]

Personally, I'd like to see something like what happened with 1.5/1.6. A new server is started for 1.12, but 1.11 remains up. Add-ons that get uploaded to 1.12 are removed from 1.11, while those that don't get updated remain. If someone uploads a broken add-on to the stable server, they are booted down to the old dev server, but not deleted entirely. This way, they are removed from the choices that visitors are offered, but are still available for those who really want them.

How would we judge what's broken and who would wear the boot?

[Insinuator]

Server admins most likely, Drankof. I'm not endorsing or criticizing Groggy Dice's idea, but it's a different issue entirely. "Functional" and "Quality" are very seperate.

As I mentioned earlier, I'd like to make Reviewing a Feature Request, but only if I can get explicit support from a developer. And perhaps some guidance on doing so via the GNU site. It's been more than a few years since I used that form.