Log in
| 1 | ||
| 1 | ||
| 1 | ||
| 1 | ||
| 1 |
I added a little rate limiting to the post submission. The little bit of spam we got often comes in the form of posting an absurd number of posts all at once. From most of the conversations I've had with others when bringing up spam it seems the majority have not seen any. That means I've done my job well. But never the less it has been happening and that is the typical format.
I did set it a bit aggressively to see if it can also help out in a few other areas. One is bot posters. Not quite spam but getting kind of close. One of the details of the spam rule here is that promotion is OK but not if users dislike it. Well, their posts have been getting upvotes and the content seems ok. It's not even promotion. So it is not spam. But it tends to be able to slide the site a little bit, so while I asked them to at least spread their posts out a bit, we'll give them a little help and make it manditory.
The last possible side benefit is culture shaping. I have mentioned in other comments that I want to encourage goatmatrix toward project-oriented content rather than just news dumping, along with other higher value content. This was a substancial portion of the content when the site was smaller, especially back when @HQDP was around and when myself, JasonCarswell, and LarrySwinger had many active projects.
Since Saidit's major migration there has been a shift in content. I don't consider the news to be bad or negative content. It's just the lowest of what I would consider good content, within a concept one could call The Pyramid of Content.
In case that image doesn't load for anyone the pyramid is basically a ranked list of high value content to low value.
Below then news you get into genuinely bad content. The news is the very bottom of what is good content. It is good content. Just the very bottom of it. The problem is that if everyone univerally upvotes everything that is good, which includes the news, then very good content can never rise above it. The score difference between good and great is very flat. At that point what determines the ratio of good content to great content is the amount they are posted. And unfortunately news is very easy to link to in bulk, and great content takes effort. A single user can decide, "I'm going to post all the news," And it will be all anyone sees. And that becomes what a site is. The very bottom of what can be considered good.
Now if GoatMatrix becomes just a news site I guess that's ok. But it's not what it was exclusively before the migration. I don't think this is a scale problem where beyond a specific threshold the culture automatically shifts. More people means more hobbies and more interests. To me the real issue is just momentum.
It's a bit like the migrant problem. A migrant never gets to spend a significant amount of time seeing how the culture operates without him. Even a good migrant runs into this problem. And so while they might come from a good culture, they don't understand the opportunity costs associated with losing the specific advantages the prior culture had. So it becomes the job of the original inhabitants to try to argue for those practices and demonstrate the advantages. And that is only possible if the new migrant hasn't already dominated the space. That doesn't mean the mirgrant isn't valuable or doesn't have things to offer. But it might be good to pause and consider what the original culture has to offer so we can figure out what the ideal hybrid looks like.
How this relates to a pretty agressive rate limit is, well, if the side consequence is it makes it harder for a single user to post a large list of news articles, and submitting them to post whenever isn't really an option because the media will have zero value very soon, then maybe that's not a bad thing. It won't kill the news. It will just make it harder to link dump it and it will leave space for other content that doesn't have ephermeral value.
Maybe it is a bit agresive and it likely can be relaxed later, but I want to give people a chance to see a different site before we lock into all news all the time and nothing else. And of course the news will still get posted. But maybe mass news posters will have to pick just the best stories for a little while. It's an experiment.
The other thing you can do to contribute to the experiment is taking on a challenge I want to issue. For every user to post at least one thing higher on the pyramid than the news. Let's see what the front page looks like. Maybe we'll like it. This is even a chance for the lurkers to step out of the shadows and post something. And not only could you post one thing higher on the pyramid. You can post many. This is what the post whenever feature excels at. If what you are posting doesn't have emphemeral value you can post it in bulk. Then we can post the news sparingly (not zero) and post higher value things in bulk instead of the other way around like happens on most sites.
Let's try it for a week and see what we think.
Update: I relaxed the alg
Thanks for the pyramid. I'll try and spend more time posting projects.
Genuinely hope this helps. I think you’ve identified the core of the problem is link sites aren’t really as fun as sites where people are sharing something they made themselves. Also memes.
I don’t know if this rule by itself will make too much of a dent, but doing things in a measured way is probably a smart call.
One suggestion I had made to AOU in the past was to implement a domain whitelist feature, wherein the user could turn off links in their feed, with whitelisted exceptions. For example, no links entirely, no links except the image hosts, no links except the news sites I’ve whitelisted. If it’s a view that can be toggled, I think it would let the users get a more interesting experience while still being able to dive into the /all everyone now and then to expand their horizons.
AOU had previously tried a domain blocking feature, but it became resource intensive if too many domains were blocked. Maybe it’s difficult to implement a domain whitelist, too, and that’s why he never did it.
Anyway, good luck.
Wow. User whitelisting and blacklisting domains. That's and idea. But side effect of the idea, not your fault, is that you've just added one more straw to an already broken camel I'm currently trying to lift off the ground. I'm wanting to eventaully implement this idea I'm calling parties which is group managed lenses for content that can be easily turned on and off. For example you could have the no-news party saved in the parties you use, and when you want to view the site that way you just click it on.
But what I keep struggling with is that the scope and way I'd do that is shifting like sand in my mind. The absolute ideal way to do it would give so many tools to a party it would be nuts. I hadden't even thought about domain white and black listing.
I guess a way to keep it sane for my sake would be to target features for users first since that's more simple, and then to slowly move features that are available to individuals to parties.
Maybe there is a reason why I'm not the first to think of this idea and if I were to try to make it I'd likely be the 50th to fail at it. You really need to let (users,parties) (whitelist,blacklist) (subs,users,domains). And what happens when multiple sources converge? Can they import each other? And ultimately it has to run on a database.
Maybe it's not really that complicated but with even new very valid and good idea the question is.. does this new consideration throw out everything I've thought about?
The parties idea sounds cool. Maybe you can roll out customization in different stages so it’s more manageable.
Global links on/off toggle. If you turn it off you just see text posts + direct image links. Same for everyone. It’s just a button on the site.
User-defined domain whitelists layer over links on/off toggle. If you turn links off you see text posts + direct image links + what you whitelisted
Ability to join and create parties.
It seems like #3 would naturally flow from the whitelist features in #2 and the already present feed customization slider.
Btw, here is where tags idea would be useful - you can introduce "meta:personal projects" tag and uprank it for front page (and downrank "news" tag - this is normal content tag, not meta). Or even better - give users sliders to change on custom settings page.
Yep. And it would also be possibly useful for the "parties" idea I want to implement. It's basically group managed lenses. A party could select that they want to see mostly content with a certain meta tag. Project party: whitelist specific users who post a lot of projects, whitelist any post with the project metatag.
The problem is figuring out how to fit that with the existing database schema.
So what's the rate limit? On Saidit it was 8 posts/day I think (recommended one, not always enforced).
So I meant to talk about the how but wrote too much about why. So the way it really works is there is a metric for if you are currently dominating the front page. So if you are over half the posts on the front page or over half the posts in the top 5 it will block. So it won't impact 90% of users. Just if you are already dominating you can't dominate more. It shouldn't take too long for others to post.
This is kind of an idea we had before. I just decided to make it happen because it seems to fit with a few things we need/want at the moment.
The "post whenever" is a bit of a drag. Any way you could whitelist users who have passed the human sniff test?