Believe it or not, you can! Check our the documentation https://coralproject.github.io/talk/ where you can use plugins to hook any any part of the commenting process.

Are you saying I can run plugins as a site user (not administrator)? That doesn't seem to be the case.

https://coralproject.github.io/talk/docs/running/plugins/

As a site admin, not a user. If you see our server side plugin api [1] you can see ways to extend the graph schema, hook into existing mutations/queries completed by the system to apply any set of rules or moderation policies you want. It would be a neat implementation if as a user, you were given the ability to create your own filters, which would certainly be possible if implemented via a plugin! But don't expect to do that on WaPo anytime soon.

[1]: https://coralproject.github.io/talk/docs/plugins/server/

Thanks for the explanation. I would be interested to help build a plugin that implements user-local moderation policies, if sites would actually install it. If that's not likely to happen, then a distributed commenting model seems more useful to me.

User-local moderation will be hamstrung by the lack of data. Sure, you can block certain users or reorder the results a little bit, but anything interesting would probably require the server shipping over huge amounts of data to a client which may never even do anything useful with it, which sites are unlikely to do.

Your best bet if you really want this feature is to implement it yourself server-side with just enough knobs to make it useful for your user-local case, and hope that other sites agree...

Individuals publishing psuedonymous moderation would provide the necessary data, provided privacy concerns can be addressed.

I.e. as I upvote your comment, I publish the info that "user XYZ upvoted this comment" which you can use in your own user-local moderation schema (at your discretion).

What we have to avoid is when Potential Employer looks for info on "John Q. Smith", they can identify which articles he upvoted.

Good site also to check out is our Guides [1]. We've collected research on user engagement into a format that should be easy for newsrooms to adopt to create better discussion.

[1]: http://guides.coralproject.net

The feature I want is, as a site user, to apply my own set of filtering and moderation criteria.

E.g. "I don't want to see comments posted by user XYZ, or containing regexp W. Any comments by user ABC, or upvoted by ABC, or containing regexp D, go to the top. I trust the moderation of site trollblocker.com"

I don't see these kinds of ideas represented in the user engagement research.

One of the features offered by Talk is the ability to ignore comments from a specific user [1]. It is designed just as you described it here. As mentioned in a previous comment, our server api's should be sufficient for you as a site administrator, to write a plugin that would allow users to do exactly what you described. The idea with Talk is that we hand control to the newsroom, so they can enable or create the plugins that they like.

[1]: See a demo GIF on https://www.washingtonpost.com/news/ask-the-post/wp/2017/09/...

This is a step in the right direction, but why hand control to the newsroom as opposed to the user?

It is the newsroom's site.

Then this is evidence that comments should not be on the site. It's not the place of the newsroom to control discussion.

Then don't comment on their site? I mean if you're looking for discussion online, there are thousands of places to go. But this is specifically about the comment sections they provide. It seems entirely reasonable that a room they provide should be subject to their rules, and one violating those rules should be shown the door.

My observation is that Mozilla Talk does nothing to advance the decentralized Web, so why is it a good thing?

Just because something doesn't advance your pet cause doesn't mean it's not good.

The decentralized web does nothing to advance universal health care, so why is it a good thing?

I'm just going to point out that Mozilla and the NSF recently offered $2M for ideas to decentralize the web.

https://blog.mozilla.org/blog/2017/06/21/2-million-prize-dec...

They never said that Talk was aimed at moving that goal further along, however. They're trying to use it instead to allow better curation tools to make comments manageable.

It seems that their goal is to help sites move away from 3rd-party comment management services such as Facebook and Disqus. Based on this discussion, I think federated (not site-owned) comments are the way to go long-term. Maybe Talk can eventually be adapted to help with decentralized comment moderation.

The thing about federated comment feeds is that these news outlets will not be incentivized to embed them directly to their website, as they lose control of the content. This problem will damage the exposure that a federated system would receive.

Better than having news outlets embed them, is having browsers support mixing federated comments with the original content. How about it, Mozilla, Brave, and Chromium? Afraid to bite the hand that feeds you?

Okay, I'll give it to you that your idea would be really damn interesting. Still, there would have to be some semblance of moderation, and as to who that responsibility will be placed on is a hard question.

Agreed. I trust the guy who runs uBlock Origin over AdBlock Plus not for any great reason, but because I vaguely remember people on this site recommending it.

I imagine people could delegate moderation responsibility to people and organizations with a strong reputation. Let's say GNU, EFF, and Larry Lessig would all publish moderation data and metadata; as well as your "friends" on various social media sites, friends-of-friends, etc.