Meta HN: Prevent new HN accounts from posting for n days?
Would, say, a 3-day waiting period before allowing posting (both new comments AND submissions) be worth trying out?
(x yesyes, i know i'm just going off of feels/vibes/etc. here)
Would, say, a 3-day waiting period before allowing posting (both new comments AND submissions) be worth trying out?
(x yesyes, i know i'm just going off of feels/vibes/etc. here)
On the other hand, it might discourage legitimate new members posting for the first time, like when they notice their work is highlighted here, or a long time lurker who feels empassioned to respond to something.
Spam doesn't seem like that big of a problem right now. Between green names, downvotes, flags, and @dang, I think we generally do a good enough job of filtering out and preventing them from floating to the top. I'd rather have a few bad ones get through than dissuade someone new who's earnestly trying to post.
> throwaway green accounts for drive-by snarky comments, spam, astro-turfing, etc
User flags on such posts are extremely helpful. They help by downranking and/or killing bad posts, but even more by providing a high-value signal for software to classify future posts. When you flag posts like that, you're not just cleaning litter off one city block, you're training the street system to deal with future litter automatically.
This only works to some extent (bad actors will always create new accounts), but the extent is considerable. If you care about HN, it's a valuable way to contribute.
(You don't need to do this for garden-variety stuff, but if you see something going above and beyond, it can help.)
I've done this many, many, many times, on all manner of issues. Dang almost always responds within an hour or so, rarely more than a day. Quite often the issue's already been dealt with, either automatically (HN's more complex than it appears), or by mod intervention. Occasionally (and increasingly since I've tightened my own criteria for reporting) I seem to be reporting unrecognised abuse. That's not because the abuse is more common, but because I'm reporting issues less frequently and after additional checks, e.g., looking for previous mod admonisments.
You can check for this past history using the "replies" endpoint, which can show where a particular profile's had direct responses from dang. For myself, e.g.:
https://news.ycombinator.com/replies?id=dredmorbius&by=dang
Keep reports brief and specific. Link directly to relevant content or profile(s) within the body. If you're reporting on content, include the content ID in the email subject as this apparently makes processing reports much easier.E.g., the content ID for dang's comment above, URL:
https://news.ycombinator.com/item?id=42570159
Is "42570159" (the digits at the end of the URL).You'll find similar contentIDs on all submissions and comments.
So I need a throwaway account. But if I have to wait 3 days, the conversation is over, and nobody will see my comment.
However, there are a number of other techniques based on behavioral analytics or IP/email reputation that could detect and block spammers. The age of the account is just another metric we use in the rule engine, but it should be used independently to avoid spoiling the experience of readers who have converted into new writers.
Not advertising, but our team has just open-sourced a user analytics platform that could help exactly with this.
HN already has a lot of anti-abuse baked in, much of it highly dynamic, and as dang's said here, largely based on member-based actions. Downvoting and flagging content really does work, even if it's occasionally slow (a few hours to accumulate enough flags to down a comment or new account).
If that's not fast enough for you, or you see a newly-emerged problem, email the mods at hn@ycombinator.com, with a clear and concise statement of the problem and links to examples. More on that in an earlier comment, including some research tips:
<https://news.ycombinator.com/item?id=42571073>
Most of the time I've reported spammy or abusive accounts they've already been marked as such within the system. Maybe one in ten seems to be novel.
I create those accounts to share those opinions to understand and see how they are received. To my surprise, 50% of the time, they are received extremely well and I really need to do a check on how the world works. The other 50% of the time, it'd go exactly as how I thought it would go.
So, for me personally, I'd be against it as I learn a lot about it.
Moreover, I've seen other throwaway accounts do the same. Sometimes these comments tend to be upvoted a lot as while it is spicy, there's some strong resonance there among the HN crows. Often times, it's also downvoted into oblivion. I feel like I can differentiatie it compared to the standard snarkiness, spam or astroturfing. I find both the upvoted and downvoted versions valuable as I appreciate candor.
just flag those comments
turn off the computer
touch some grass