This is something I've been thinking a little bit about recently, too, although I don't have many ideas or much of a plan yet.

Do you have thoughts on how to scale this or what the feature looks like a few iterations out?

I think the approach in D13818 probably isn't quite right, since it saves who can administrate an answer in the database. If I have the wrong setting and someone writes a really bad answer, there's no way I can fix the problem. Is that something you have a plan for?

An alternate approach would be to just have a "Can Manage Answers" policy, which you don't save on each answer and just look up every time. So instead of doing return $this->getEditPolicy(), you would just look up the configured policy, like this:

return $app->getPolicy(PonderAnswerCanManageAnswersCapability::CAPABILITY);

Then if I change that from "Admins" to "Just Chad" in Applications → Ponder, the change is effective immediately and retroactively.

My major concern with this is that I don't see a very good way to scale it if there are a lot of questions/answers. Maybe we don't need one, or we don't need a plan right now and can deal with this issue later, but I imagine users may want to let Matt and Jane administrate different sections of questions. This might reasonably never come up, but I don't really see any great options if it does, since we don't have much of a way to separate questions into sections.

Maybe this isn't really an issue. I think both Quora and StackOverflow give administrators general-purpose powers, but both primarily rely on community moderation, and I don't think moderation on either site is especially high-quality or consistent, nor does either site have much of an editorial voice. I personally view these as fairly strong negatives, and consider editorial voice really valuable.

In contrast, a site like Reddit gives moderation powers within sub-communities, and a lot of forum software works like this too. If we were building a forum, I think this would be pretty obvious/straightforward, but it's not clear to me that putting hierarchy into Ponder is very valuable/useful/worthwhile.

Marcel has a forum which uses a sort of twist on traditional subforums: each topic is posted to one or more "tags", and moderators have power within a tag. So tags are a bit like sub-forums or sub-reddits, except that topics can easily/automatically cross-post to a bunch of them. When you post to a tag, administrators on that tag get some power over your topic (in exchange, you get distribution to users who browse that tag). This sort of "tag cloud" model seems to work pretty well, but I think the tag moderators don't have much editorial voice either: they can mostly just kick your topic out of their tag, or maybe sticky it or do a couple of other things, a lot like if Reddit let you cross-post automatically.

I think allowing there to be a very strong editorial voice is generally a good thing. I think Quora/StackOverflow have a lot of content which doesn't violate any rules, but which also shouldn't really exist. For example, the first question in my Quora feed is this:

https://www.quora.com/Is-Mark-Zuckerberg-a-bad-person-as-depicted-in-the-movie-The-Social-Network

The #1 answer is from Jimmy Wales, who personally knows Zuckerberg. It actually answers the question.

This is followed by 21 answers which I'd claim are all garbage and shouldn't exist. From skimming it, it looks like none of them actually know Zuckerberg or are citing real sources or actually answering the question. They're just telling stories, guessing, ranting, promoting themselves, copy/pasting articles, etc. By the bottom, we're getting stuff like this:

Most people are neither good nor bad. Just a combination of both.

Following this are 19 additional collapsed answers, which are even worse.

So we have 41 total answers, roughly ordered by quality, and about half of them are good enough to exist. I think this is generally consistent with community moderation: community moderation can filter out the total garbage, but it has to err heavily on the side of caution to avoid censorship, brigading and drama, so it can't have much editorial voice and generally can't filter out stuff that's low-quality / not useful but doesn't actually violate any rules.

I'd expect a similar result from community moderation on StackOverflow too: you'll have a couple of helpful answers, a big chunk that are useless but don't actually violate any rules, and then a chunk that are such total garbage that community moderation is effective in getting rid of them.

I'd ideally like to see the editorial bar set much higher, and for those 20 bad-but-not-complete-garbage answers to generally get tossed, leaving about 1 actual answer on that question. This is much less relevant in private installs of Ponder than it is for public sites like Quora and StackOverflow since a Ponder question probably won't get 40 answers from people wanting to tell a story about the time they watched a movie, but it's an editorial bar which I think isn't achievable through general community moderation and requires editorial voice. I'd ideally like to set the tools up so that empowering this editorial voice is possible.

Anyway, just some vague thoughts about this.

I think it would be reasonable to just give this power to all admins for now, I just don't see a super clear way forward if that proves insufficiently fine-graind.

I pretty much agree with all of that. The fine grained approach can I think be punted on, I don't expect to compete with Discourse. Basically if we ever had "PonderTopic", we could split moderators up per topic or do the TagCloud thing.

Mostly it's hard to imagine any install getting more than 3-4 questions in a day, today.