For sites without account creation, sure, you might be losing a ton of good content... but it won't matter. The spam, harassment, trolling, illegal content, and the pinning of responsibility on the platform owner will almost certainly get you shut down. The problem of how to scale moderation of human communication is unsolved. Until it gets solved, accounts--really, identity and authentication--are not really optional... at least not on any website I would choose to run.

One way around it in niche contexts is to run everything through a determined moderator.

It seems inactive now but Edward Tufte's forum was a great example of this – https://www.edwardtufte.com/bboard – with every single post or comment being made via a basic HTML form and Tufte would then approve everything manually. This resulted in nothing but high quality (and often authoritative) discussions as you knew Tufte was directing it. Indeed, the forum even discussed the topic of content moderation once and Hacker News popped up: https://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=...

This can't really work at scale, but then would I prefer to frequent a handful of gigantic forums or many Tufte sized hyperspecialized forums..? In any technical field, the latter.

The Online Photographer is another excellent example of this approach (albeit with blog comments rather than a forum):

https://theonlinephotographer.typepad.com/the_online_photogr...

The interesting bit in the article for me was that if the account was required, and the lame-content posters were gone, the good-content posters also posted less (seemingly, because someone else wasn't wrong on the internet.)

That was indeed interesting, that "for every bad contribution prevented, seven good ones never happened".

However, while it is easy to measure, I'm not convinced that "number of good contributions" is a valid proxy for "quality of discussion" or "quality of content".

For example, some of those good contributions might have been repairing the damage made by bad ones. And on a less heated/combative forum, users may feel comfortable consolidating multiple small contributions into fewer but more well-thought out contributions. And perhaps even the good contributions that were prevented were (while good) on average lower quality than the ones that remained.

For a community that really valued anonymous posting, couldn't you let members vouch for anonymous posts before making them visible? If no one vouches then they disappear.

Like, hacker news almost allows anonymous posting because the account creation is so simple.

Maybe, but "a community that really valued anonymous posting" is likely to be defined by this value. The first "problem" is that likely use cases are users/comments/topics that aren't allowed elsewhere. So.. that community will tend to be a haven for whatever content other sites have actively ejected.

That doesn't make it impossible to have anonymous accounts, but it is a high tradeoff decision and likely a defining one.

I think it's more applicable for sites that have somewhat more objective and/or legible goals. Wiki-ish sites, for example. If there's a lot of free form discussion... anonymity is probably a bad decision for the site owner.

One of the amazing things about Reddit is how different communities compete with each other.

Allowing individual subreddits to enable anonymous posting lets the market decide if the pros outweigh the cons. (If the subreddit has a history of bad moderation, ban the subreddit)

I feel like giving communities more options in customizing the algorithm that dictates the discussion is powerful.

Yeah. Reddit's structure gives us a lot of insight.

This will end up in a situation where some people vouch for illegal stuff because they like it, or "for the lulz".

For example there were multiple "name a product" contests raided by trolls- because nobody moderated the inputs.

HN has its own problems: the "flag" feature is abused by brigaders to censor content.

You can ban the lulzers who vouch for stuff you don't want.

If its an anonymous community you can't they will just switch to a different ip on their vpn etc unless you can deanonomise them

The proposed community still has non-anonymous members who could accrue good standing, it just also allows anonymous posting.

Hacker News allows it (how else are you gonna get that hot FAANG goss?) but it's actively discouraged:

"Throwaway accounts are ok for sensitive information, but please don't create accounts routinely. HN is a community—users should have an identity that others can relate to."

https://news.ycombinator.com/newsguidelines.html

HN has the flagging system, votes, regulars, and the mods.. and it scales very well they are always on top of it.

Gab, the free speech social network, deals with this topic a lot... and likewise does moderation with a flagging system, volunteers, group mods, and site mods. The community is far larger, but our policies are much simpler to compensate.

The first part of solving this problem may lie in tying identity with micro payments to make and read updates/comments on content. We would do this by implementing the 402 HTTP code stub using the Lightning Network. There already exists a proxy to implement this: https://github.com/lightninglabs/aperture

The second part of implementing this is to increase the cost of producing bad content for a particular identity. Using a proof of stake type system based on the identity and content, not the coin itself, might allow decent scaling of moderation.

> work by selecting validators in proportion to their quantity of holdings

In this particular case, the comments on the content (holdings) become the selecting validators. The more high quality validations through consensus of other identities, the cheaper it becomes to comment as a particular identity. If the identity becomes a bad actor later, subsequent consensus from other identities increases the cost of that identity posting new content related (through identity) to their older, higher quality content. In a way, the content an identity produces becomes their "ego" or "account".

But, by using crypto identities, no "account" is needed. A tag may be added to the content in a particular system to indicate which transaction and identity was responsible for paying to add that content. The system remains secure for privacy of participating individuals, other than what they leak through posting their thoughts to a public forum.

The author of this post (Mako) is standing next to me right now, and was initially excited that he could comment here on HN without creating an account. But of course the "add comment" form directs to a signup page!

He can comment on 4chan without an account.

Don't worry the feds keep an account for you when your post is intercepted by cloudflare.

Consider the same theory applied to video game modding tools. In the 90s and early 2000's the freedom to work outside the scope of the authorized system created plenty of low quality maps and mods, but it also created things like Counter Strike, DOTA 2, Galactic Conquest (now Battlefront).

When is the last time good content was made in gaming? All the mods that were made with passion are now soulless DLC and pop culture referencing micro transactions. Everything is just remakes because no one has the courage to innovate. Anonymity is an inherent good. If you don't like your anonymous peers, then perhaps gatekeeping is also good too.

There are other reasons for the dearth of mods nowdays. The PC is no longer such a dominant platform so many people who might be persuaded to try modding are playing on a device unsuited to it like a game console or phone, a pure consumption device. Another issue is production value-as games get better the bar gets raised higher and higher for content. Compare the effort required to make a SC1 map to making an SCII map, it's night and day. Finally, modding is no longer the only way to scratch the itch of game making-starting with Flash and followed Unity, Unreal and now Godot we have an embarrassment of platforms for making your own game that you will own. This scratches the same itch for people.

We just took Guitar Hero II and upgraded it for its' 15th anniversary to add a lot of quality-of-life features for players (and since it's still for PS2, it's not a "remake" or "remaster", but a real upgrade). Any "pop culture" we reference is either music or GH community related. The other two members on the team especially really carried the project.

https://gh2-deluxe.neocities.org/

Autobattlers are probably the most recent innovation, funnily enough coming out as a mod of DOTA.

And DOTA was just a custom Warcraft 3 map.

And then there's Minecraft.

7:1 decrease in positive:negative contributions seen when 137 Fandom sites started requiring accounts.

The negative contributions seem to catalyze an outsized positive response.

Reflects the experience of Wikipedia as well.

Apparently crowds know how to deal with themselves.

There's a pretty big difference between wiki-first sites and discussion-first sites, though.

Exactly. You can't, imo, usefully abstract login/anonymity from the context of what the site is. Is it a chatty forum, like hn? A wiki has somewhat objective or predefined goals. That legibility helps frame things and flame wars need to resolve somehow.

When a discussion thread on a forum erupts into flamewar, we just move on. No resolution is necessary except perhaps preventing the next flame war.

This (from the article) was unexpected -- Surprisingly, the cost includes “lost” contributions from community members who had accounts already, but whose activity appears to have been catalyzed by the (often low quality) contributions from those without accounts.

So the crappy hot takes that cause the "quality members" respond to with a quality post are a net value to the content quality?

I'm not certain if you're being tongue-in-cheek or not, but I would think that often this would be positive net-value to content quality. If it takes a troll or a bot to stimulate quality discussion... so be it, I suppose.

sounds like a "The fastest way to get a good answer on Stack Overflow is to post a wrong answer with a second account" type of thing.

The fastest way to get the answer in a organization is to send a email/slack with a wildly inaccurate answer to the question no one wants to answer/respond to. It is even more effective when you attribute the wrong person to it. This is a anti-pattern but it does work.

"Hey everyone, I just heard the Phoenix Project that Jim is leading is going to be completed next week and will be under budget by $5m. I will go ahead and report that up to the VP in my 1x1 today"

You may look stupid but people will get really emotional and give you the real answer you need very quickly.

Because people like to contradict someone more than they like to help someone.

This is also a standard LEO interrogation technique.

"Well, son, your buddy just told us you killed her with a hammer."

"NO! It was an axe! I mean... oh, damn.."

I too found that surprising.

My assumption is that a low quality entry spurs some one to action to edit it and since they are already there they mind as well put effort in to improving it. I still need to read the paper though.

I believe this study was done in the context of wiki entries (Fandom) so I’m not sure if this does, and it may, generalize to conversational forum posts as I think you’re saying.

Pretty much this: https://xkcd.com/386/

> Surprisingly, the cost includes “lost” contributions from community members who had accounts already, but whose activity appears to have been catalyzed by the (often low quality) contributions from those without accounts.

The best way to get high-quality feedback is to say something stupid and have the expert correct you.

1. That feedback is not really as high quality as asking an insightful question.

2. Doing this burns trust, especially if done deliberately.

Depends on your audience.

I can point to a handful of low-effort journos whose whole schtick is asking dumb questions on Twitter so that the would-be know-it-alls can post "well, actually" to them in the replies.

By virtue of the engagement numbers they clearly love it, and come back for more slop every day.

#1 Requires some knowledge to do, which can be hard. I do agree that #2 is very true and indeed what I mean by bein stupid is ”exposing your own stupidity/lack of knowledge” rather than playing stupid.

I get that sites want you to create accounts to write comments and create posts. But back in the day, I found not one forum, but many, that wanted me to create an account just to look at the forum posts. It was great when a google search took me to the site, and just to look if the given site would help in my problem, I would need to register.

I'm trying to sound out a generalization of the idea, where liberaility creates the opportunity for outlier upside, provided it survives long enough to recieve it, and assumes outlier downside is somehow mitigated. There's a gamblers ruin analogy there, and markets staying irrational longer than you stay solvent variation in there somewhere. It reminds me a lot of what I've heard about VC portfolios as well, where the incentives all point to just getting exposure, and YC's startup school and other accelerators are a way to do a startup without the friction of making a company. Maybe it's related to just applying the Stopping Problem to anything with an exponential distribution?

Generalizing this could be really useful, even if it creates the 4chan of venture capital.

In this day and age, ad companies are partnering up with sites in order to be able to use the data related to account creation to help bypass people's defenses against their spying.

As a result, I've become very strongly resistant to creating accounts. I'll do it if for some reason I absolutely have to, but the vast majority of the time, I'll just pass on the site/service instead.

Funny/coincidental that the paper is paywalled and needs an account, thanks for the preprint!

What’s the best compromise? Fandom sites are a category that doesn’t not represent all site contributions. The high value contributions are changed more likely due to the volume, scaling will have different properties which the paper mentions in group size and how Wikipedia has bots to detect it, but there is no reason why a fandom site can’t have API tools to do that as well.

I liked the mention of a captcha as well, to prevent the lower quality spamming, which can be removed through reputation through IP when they make consistent high quality contributions. Wikipedia has shadow accounts made with IP, and remove contributions from low quality additions by filtering them, so the best compromise to me seems to require accounts that don’t look like accounts IP, device recognition, and identification without official registration.

On websites with required accounts, I don’t see any evidence that accounts contribute to quality, it’s more important to have a community that cares for quality, many sites need registration and just have large amounts of single use accounts that required registration, and sites like Reddit will tie posts to a username and often have an ego attached with their site score for a quick unfair value judgement which is popularity based, the judgement on a post is often tied with history, rather than judging the post by its own merits, while wiki contributions will not usually be knowingly tied to certain users.