[epivosism]

I recently have been thinking about UGC site in general, and it seems like they generally go two ways:

1. Allow most legal content to be uploaded and control distribution with algos

YouTube, Roblox, Twitter, Tinder, Flickr, Insta, FB, TikTok.

This lets users practice and test things without risk to their work or account.

2. Sites that try to "keep the db clean" by nuking stuff they think is "bad" by some criteria.

Sites like this: Wikipedia, most big subreddits

Type 2 sites can be unsustainable because they tend to make new users feel judged, and don't give them the chance to iterate and improve their work until it's more ready to be shared and useful to a broader audience. You just find your content nuked, or removed from the subreddit, or downvoted a ton, often with a dismissive or aggressive comment. This is NOT the way to grow and survive as a company over a long period of time

Obviously, there is no necessity to keep the db full of only high quality items. As the scope and number of niches a site covers, it's not possible to maintain that. On the other hand, using algos lets you do interactive tests with content, directly testing against various audiences to see if they like it, without having to do editorial work yourself.

Of course, there has to be some limit - articles for every pokemon, or every version of every pokemon, etc at some point it does get too far. The thing for me is coming in and seeing your content completely deleted.

[starkparker]

> 2. Sites that try to "keep the db clean" by nuking stuff they think is "bad" by some criteria.

The problem on Wikipedia is less about having standards and more about having changing interpretations of fundamental standards re-classify large swaths of previously acceptable content as unacceptable.

The trend away from subject-specific notability and sourcing guidelines to applying one notability guideline generically to every subject, regardless of the intent in doing so, mostly just gives editors who like to delete things a whole new buffet of articles to tear through.

[epivosism]

I'd definitely agree that changes like that intensify the issue!

For me as an amateur contributor, though, even the normal experience was tough - suddenly I returned to find my content removed, with a short, aggressive note, left by what wasn't a committee, or a jury, but rather a single, annoyed and tired individual, who in his seniority was somehow justified in unilaterally acting alone to delete my work, without the careful attention to detail in documents, wording, and formality which helps make real-world justice palatable. The carelessness with which he wrote made it clear that at that site, I was to be considered a person of very little value at all.

To speak more clearly, that was the social/psychological effect of it - which left me without much motivation to continue! And I think that's a wide effect - if you compare a typical 1st-time wikipedia editing experience to the more successful UGC sites, it's pretty clear they all make some effort to shield users from that (if they can) and let them feel successful, even if they aren't yet. (I know from the inside of one of these places that users get mad at for removing their creations for copyright or other reasons, all of us on the inside hated potentially frustrating a creator, and admired them for trying over and over to create something.)

Moreover, from what I've heard, there used to be more freedom on wikipedia, so the argument that "it's always been that way" might not even work.

[ryandrake]

Using those definitions there are really no type 1 sites except maybe 4chan, which itself even lightly moderates. All the other sites you mention heavily moderate and most of them automate that moderation.

You really can’t have a site that’s a free-for-all because of spammers, griefers, racists, and other various forms of jerks.

[epivosism]

Yes, I didn't break them out but I do consider some types of basic filtering to be legitimate because they're both common across cultures, and their limits can be described fairly precisely. These are things such as prohibitions on political, ethnic or religious conflict, criminal activities, nudity, and of course, copyright.

Actually, treating censorship of those things as the default global level of censorship makes the difference between restrictive sites like Wikipedia and the rest of the Internet even more stark.

It's almost unimaginable that YouTube would remove a creators content simply because they thought the topic was irrelevant, or the quality too low. So I agree we can't have a true free for all online. But the distinction between the large group of UGC sites that really try to cultivate and protect their creators from those that don't is reflected in the popularity and growth rates of them all.

[itishappy]

I may be missing your point, but Wikipedia and most big subreddits have proven quite effective at growing and surviving.

I'd also suggest a major distinction between type 1 and type 2 sites are a focus on creation vs consumption. There's a lot more consumers than producers and, depending on your goals, it might improve the experience for more users by deleting the content of some.

[devmor]

>most big subreddits have proven quite effective at growing and surviving.

That depends on your criteria. If popularity and user engagement are the only important metrics, this is absolutely true.

If however, clarity of purpose and effective moderation are important, I would strongly disagree. From my experience, most big subreddits that used to fulfill a certain niche have devolved into primarily meta-posting and stealth (or not stealth at all) advertisement.

Of the exceptions to the above, many now just fill the exact same purpose. There are around 10 extremely large subreddits that regularly make the front page that are essentially just "look at this picture of something I have/something I saw" with no real boundary between them.

[epivosism]

I have heard the editor count at Wikipedia has been shrinking for years. Might have been wrong though

[yorwba]

Monthly active editor count peaked in 2007 and then started declining, but has been quite stable since about 2014: https://stats.wikimedia.org/#/en.wikipedia.org/contributing/...

[epivosism]

That matches what I would expect. My normal continuation would be to look at the growth from 2014 til now in global internet users, and at the competitors like youtube, insta, etc. I'd expect to find that many of these sites grew 100 or 1000x, and that the global online population whose device power and internet service levels surpass some threshold has also grown by at 10x since then.

Those are all assumptions and memories from working during that time period - possibly wrong, would look up if you disagree. But if you don't, then my conclusion would be that staying flat during a period like that - when you could have grown and supported a deep, new, persuasive trend of literacy, fact-based reasoning and understanding, education and compromise - is a huge failure.

[yorwba]

There were 2.79 billion internet users in 2014 https://ourworldindata.org/grapher/number-of-internet-users?... so the number couldn't have grown 10x because there are no 27.9 billion humans yet. The latest data is from 2020, so maybe it's at least 2x what it was in 2014.

Most of those new internet users are in Asia and do not use English as their primary language. If you look at editor counts of other Wikipedia editions; Hindi https://stats.wikimedia.org/#/hi.wikipedia.org/contributing/... , Indonesian https://stats.wikimedia.org/#/id.wikipedia.org/contributing/... , Thai https://stats.wikimedia.org/#/th.wikipedia.org/contributing/... , Persian https://stats.wikimedia.org/#/fa.wikipedia.org/contributing/... kept growing after the English Wikipedia saturated.

[epivosism]

Another type 2 site I forgot: stackoverflow

And another type 1: github

[kiba]

Moderation is a must or else there will be rampant misinformation and hate speech.

Reddit is fine but it's not really operable as a business enterprise unless you are fine with making a small tiny profit over long period of time.

Subreddits are a self solving problem over time. They grow big and shrink on their own merits and damage is self contained.