Quality measure in mini app picker

[Simon]

Currently, the apps in the mini app picker are ordered by recently added and there are many apps to choose from.

It’s impossible to see what apps are good at a first glance from description alone. This leads to disappointment in the app and in DC mini apps in general.

Possible Measures to improve it:

A. A Badge for tested quality apps that follow some standards like having collaboration when you’d expect it and having nearly no bugs.

B. Have a featured apps tab that only shows presentable mini apps, to get a good first impression by new users

C. Rating system with stars

• server for sth. like this is not hard to do, see readme.md · master · Banana Hackers / simple-ratings-server · GitLab
• review comments are harder, since they require moderation. To keep moderation effort low they could be shown only once they were approved.

D. allow users to have a list of their favorite apps (this was already suggested in another post, but discourse search is so bad that I can’t find it)

Suggested Metrics for apps

• Does the core functionality work?
• If you already see it was low effort by many bugs, then that is bad.
• Are webxdc features used where you would expect it?
  • chess app should have multiplayer, editor, todo lists and collaboration tool should have collaboration
  • tool apps without persistent state, like some unit converter on the other hand don’t need collaboration
• Is UX clear to nice to use?
• is the design consistent?

[Simon]

The AI vibe-coded apps might add to the problem, but remember there were low effort apps long before vibe-coding even was a thing.

[WofWca]

@Simon

C. Rating system with stars

I suggested the star system in this post WebXDC app store idea: GitHub search wrapper page. Good news: yours and mine apps take the top 2.

[ian]

I think hosting a user-controlled rating system for this is overkill. Displaying the number of stars on the repository will bias the results based on the size of the forge and the number of developers using the given app (as normal users rarely have an account on such forge).

It would be perfectly sensible if the curator of the given app store listing just added static ratings themselves. A “Featured” tab is just a binarization of this, but it would work just as well to give 1-5 stars in the same way.

The ratings of the curator could be stored in the same structure where other curated tags could be also added as mentioned here:

Request for annotating webxdc with categories and labels

We could add further structured metadata items to the manifest of each app. It could be either stored in the original zip, or separately by the store. The store web app frontend could provide filters for these metadata. Combinations of filters should be invoked by deep linking using the URL fragment (URI anchor). Some of these may be mutually exclusive categories, while others may be overlapping tags. Most of these could be fairly objective, while others would be somewhat subjective (small vs. l…

As a more objective alternative instead of the star rating, the curator could tag an app with each aspect that you mention above such as whether there remain any known, but unfixed bugs or not (even list them in a small textual note?). I have already proposed tags to mark the scale of interactivity in a given app.

I can’t provide guidance about how to note UX & design issues, but I guess the more grave ones could be handled as bugs.