5 ms·
I was about to suggest a feature where we could collapse comment replies below a certain vote threshold but then decided against it. It's better to leave the re
by hliyan 6y ago
I was about to suggest a feature where we could collapse comment replies below a certain vote threshold but then decided against it. It's better to leave the responsibility of scanning through and deciding what to read on the reader themselves, rather than an algorithm. Which is where problems usually begins. At times it's a bit of a pain scanning and scrolling through reams of comments trying to find things I find valuable, but it's better than the alternative, I feel.
- tpxl 6y agoHaving to expand every single comment in reddit threads is infuriating. The HN UI is just the right amount of information dense and readable for me.
- aldanor 6y agoIt's ok because HN is much smaller; on Reddit, there are posts with 20k+ comments - for those, thread collapsing kinda makes sense (which doesn't validate collapsing comment threads everywhere else though).
- 411111111111111 6y agoComments already get marked as dead/hidden if there are too many downvotes I thought? You actually have to go into your profile and enable a checkpoint just to see them at all.
- jsinai 6y agoI agreed that it’s better than the alternative. Also the alternative has the risk of amplifying echo chamber viewpoints. And then there is the fundamental of how new would comments break the barrier of getting upvoted in the first place?
- TeMPOraL 6y agoPerhaps a small visual indicator that's proportional to upvotes would be helpful, but other than that, I too wouldn't collapse anything automatically. The thing modern web UX trends forget is that human visual system is amazing at pattern-matching and filtering. It's literally what it does all the time - subconsciously sifting through large amounts of data and focusing on things of interest. As long as there are patterns in the data the brain can use to slice and dice it, you can dump a lot of data on people with no confusion or loss of effectiveness.