Fantastic! Btw, check out codeproject's rep system if you've never seen it; seems like a similar idea might tie into what you're planning here...

For tech/scentific sites, it definitely doesn't make much sense to make a tight coupling between moderator suitability and domain expertise. If we can manage to separate those, that would be great! Rep should probably(?) reflect domain expertise, nothing else.

I don't understand the linkage between suggested edits and tag creation. Being good at proof-reading and reformatting text is a different skill to categorisation. Worse than that, not being able to create a tag can block someone from being able to post a question, especially in a newish site.

Separately, one of the lessons I took from SE was that not allowing new users to comment has negative value. Either they have to abuse an answer to leave a comment, or valid critiques get lost because people don't want to (or remember to) come back after gaining rep from answers. IMO it would be better to limit new users to one comment per day plus a rebuttal to any addressed at them.

@Lundin We'll need to discuss what rep should become (or how it should be replaced) separately. We'll start a discussion about this in a few days including some suggestions what user cards should look like.

@PeterTaylor you make a good point about tag management. Our idea was "editing posts" < "editing tags" (and both with "editing"), but this can be changed. We can require successful posts, suggested edits and flags as possible criteria for abilities. What would you suggest (currently it could be called: "edits: high", where high is a qualifier for the amount of edits needed)

And about comments: The rate limits can be edited by per-site. So we can react to sites where comments-as-answers are a problem. I'd leave the limit at 0 for now and revisit this later when we have actual data to verify.

Editing tags as an extension to editing posts makes some sense, with a possible extension that people with lots of good content in a tag can edit the description of that tag. For tag creation, in a new site I think free-for-all might be the best option; in a mature site, maybe something like having good content in N tags, which suggests that you are familiar with the site's ontology? That would extend back to new sites by setting N=0.

Much of this sounds reasonable, but it seems like the metric of generally writing well-received answers is getting lost except for the first ability you mention. Someone with lots of good answers is going to be familiar with the site and its norms. The should lower other thresholds to get specific abilities. In other words, abilities should be based on certain actions and knowing the site well, with some squishiness between the two that can be set per site.

For example, someone that has read and answered a lot of questions should know how to tell good from bad questions. In fact, they are probably better at it than someone that had a few close/reopen/delete flags accepted.

@OlinLathrop "it seems like the metric of generally writing well-received answers is getting lost except for the first ability you mention" ... 3 of the 5 unlockable abilities rely partially on well received posts though? (Participate Generally, Vote on Holds, Curate)

"For example, someone that has read and answered a lot of questions should know how to tell good from bad questions. In fact, they are probably better at it than someone that had a few close/reopen/delete flags accepted." On the other hand, someone who knows how to tell good from bad questions should know when to flag and thus have more of their flags accepted. And if someone has their flags accepted, they probably know what they're doing, regardless of how active they are in posting.

@Moshi Right. I'm suggesting a squishy metric trading off the two. For example, you can get the ability with high accepted flag count and low contribution score, low accepted flag count and high contribution score, and combinations in between. The four limits should be settable by site.

While it's still very early, I think this might scale better than SO's system, I always thought it was very silly that I could never get the ability to edit posts without review even if I had hundreds or thousands of approved suggested edits. One thing to note is, we might not really know how good the edit privileges are until posts have a chance to age and new users start joining that post lazily formatted content.

... also I think SO's need for edits comes partially from the "never delete an answer" policy, a lot of the posts I would like to edit were technically correct but deeply flawed. The other kind are posts that are outdated due to dead links. Hypothetically if reviewers had domain knowledge I would also suggest edits to code to fix memory leaks, and other things that I would not normally think would require a new answer.

...and a third thing, collecting answer and question votes on here might be a lot easier than on SO, I vaguely remember people saying that votes used to be a lot more common on SO when the site was small (more attention to each post), and plus there's tons of subjects that have zero content at all. Though if this site optimizes for quality contribution instead of "everyone should post their homework here" then maybe the decline in votes can be avoided, I hope so.

Something to consider: Discourse uses reading stats to grant trust levels. Fundamentally it's about encouraging people to understand your community before jumping in with both feet.

@JonEricson yeah, we used Discourse prior to this Meta site for coordination and the current Abilities system is heavily inspired by them. We currently don't track or intend to track reading and page visits. Also AFAICT, reading time generally requires JS to be active, which we intend not to demand for basic site usage, to which I'd count Abilities.

@luap42 "And when you are asking where the sweet technical details are; I’ll write another blog post in a few days about how abilities are designed." So, where is it? :)

This is a question and answer site - this is why folk come here. Reputation/scoring should be dominated by that fact. It sounds like what is being proposed appears to be missing the point or maybe I just don't understand what an abilities system is.

@Andyaka I don't get what you want to say? The Abilities system is based on the idea that we should give you the tools you have shown that you can be trusted to use them. Not everyone who's (for example) good at spotting off-topic posts is also good at editing posts.