News for hackers. Stuff that matters.
Re:Orion – New WebKit Browser by dang on Wednesday June 01, 2022 @08:52PM (#31590037)
Ok, we'll move the comments thither and re-up that thread. Thanks!
Re:Orion – New WebKit Browser by beefman on Thursday June 02, 2022 @01:24AM (#31591644)
These are not the same article. I don't care at all about Kagi, but I care very much about Orion. Except I only know this from the other Orion article (the one now marked "dupe"), which shows that it has native tree-style tabs.
Re:Orion – New WebKit Browser by pvg on Thursday June 02, 2022 @09:36AM (#31594693)
They are effectively the same story, the dupes go by story sameyness rather than article identity.
Re:Orion – New WebKit Browser by beefman on Thursday June 02, 2022 @12:02PM (#31596534)
Right. That's the policy that's bad.
Re:Orion – New WebKit Browser by pvg on Thursday June 02, 2022 @12:52PM (#31597362)
You can try to come up with a better one but one based on identity runs up against the constraint of there being only 30 spots on the front page and the site's charter of interestingness.
Re:Orion – New WebKit Browser by beefman on Thursday June 02, 2022 @01:47PM (#31598189)
The better policy is simple: the dupe system should only be used for identical articles. There is no harm in allowing two threads about the "same story" to ride out the front page nor is there any reason to combine their comments.

Though unforeseen consequences of this suggested policy are possible I think they're unlikely and I don't believe anyone can reliably predict them without trying it.

As dang likes to point out, internet discourse is mostly stateless. Combining duplicates that aren't really duplicates does nothing to improve this situation.

Re:Orion – New WebKit Browser by pvg on Thursday June 02, 2022 @02:22PM (#31598637)
There is no harm

Two (or what your suggestion implies, N) nearly identical interesting things on the front page is less interesting than two different interesting things on the front page.

This is, in fact, a problem that afflicted HN more or less right off the bat as it does any crowdsourced link aggregator - Erlang days, etc. It is, as you say, entirely predictable and crushes interestingness.