Hacker News Duplicate Detection Issues
Duplicate URLs are slipping through the cracks on Hacker News and hitting the front page, which suggests the deduplication logic is failing. Normally, HN redirects a second submission of the same link back to the original thread to keep discussions centralized, but that's clearly not happening consistently right now.
From a technical standpoint, this is usually handled by a simple hash check of the normalized URL. If these are getting through, it's either a regression in their submission pipeline or a weird edge case with how they're handling URL normalization (like trailing slashes or tracking parameters) that's bypassing the filter. It's a minor annoyance, but it fragments the conversation when two different threads for the same link start gaining traction simultaneously.
Next
Image-to-Prompt: A Complete Guide to AI Art →
I've spotted several instances of this recently:
- Whetuu (Zig prompt): Two separate threads for the same GitHub repo.
- Billionaires/Innovation piece: Multiple entries for the same Elysian Press article.
- NYC Infrastructure blog: Duplicate posts for the same Practical Engineering link.
From a technical standpoint, this is usually handled by a simple hash check of the normalized URL. If these are getting through, it's either a regression in their submission pipeline or a weird edge case with how they're handling URL normalization (like trailing slashes or tracking parameters) that's bypassing the filter. It's a minor annoyance, but it fragments the conversation when two different threads for the same link start gaining traction simultaneously.
All Replies （3）
D
Drew36
Advanced
11h ago
Wonder if it's a caching lag or if they changed how they handle trailing slashes.
0
S
Wait, so we're just playing a game of "post it again and hope for a miracle" this time? Because half the time both posts just end up shouting into the void anyway. This place moves so fast that a link is basically ancient history after three hours.
0
R
Probably happens more when users tweak the URL parameters to bypass the filter.
0