1. 9

Lately, I’ve been searching through site history a lot. This works well generally, but I e.g. just read a history of APL and wanted to supply a review comment. Unfortunately, it was posted some months ago and received no discussion. We could obviate the need for reposts (and e.g. the repost checker) by allowing further dated responses.

This could effectively change the site data model, where a post collects all opinions/discussions about a particular piece/URL. I’m not sure of the overall consequences, but they initially seem rather interesting (and positive). In IRC, some people said they don’t think this is the place for canonical discussions though, in which case: When can we repost?

Positives:

  • obviate reposts
  • change data model, now post is complete collection of discussion on a piece

Negatives:

  • makes archives dynamic, requiring technical/moderation changes
  • could notify original thread poster (if they chose), a minor form of why people historically hated necros (but fixable)
  1.  

    1. 9

      It’s fine to repost a submission after 6 months. I don’t find that kind of duplication a huge issue.

      I’m however not familiar with the rationale for disabling commenting on submissions after a certain amount of time. Performance issue? Spam mitigation?

      1. 5

        It might be spam mitigation. There’s a blog I read that’s been active for 25+ years, and still allows comments on old posts. Occasionally an old post will get spammed with links in, I think, an attempt to boost (for lack of a better term) page rank for the links, and by targeting old posts, the spammers hope the links go unnoticed by anyone but the webbots.

        1. 3

          It’s fine to repost a submission after 6 months.

          In practice, however, I find that the site discourages reposts. For example, I recently wanted to submit Conventions for Command Line Options. It was posted here four years ago and received only one comment.

          When I tried to post it again, I saw this note demanding I justify my resubmission. (After I saw the note, I gave up on the resubmission.)

          What has changed or warrants new discussion?

          (Not just that you think it’s still interesting or read it for the first time.)

          From one point of view, nothing has changed and nothing necessarily warrants new discussion. (Command line flag handling hasn’t changed recently in some novel way..) Also, it is true that I recently read the article for the first time. From another point of view, when it was first submitted, the post received essentially no discussion. Also, I’m interested in the topic. Am I supposed to write that? What do other people do in such a case?

        2. 4

          Sometimes the submission or the discussion in comments contains a niche prediction with a horizon of a year or two; a later update on how it worked out for archive-searchers currently cannot be posted. Or even without predictions sometimes «how it worked out» is not something trivial to look up.

          Spam could probably be reduced by «participated in the same discussion before or has well-received recent comments» (see mods-and-owner-only «comments on non-own submissions with above-average score» calculation in the profile for the latter).

          However, what warrants a fresh run in the «newest» and what doesn’t is a good question.

          (Maybe allow bumping to everyone with some low carma cut-off if older than a year, and allow non-bumping updates under old-participation or a higher recent-good-standing cut-off? But if bumping is not a fresh discussion with a backlink but moving up the old discussion into the newest again, this is obviously a larger change than adding necroposting conditions)

          1. 3

            Interesting. I always felt having 90 days to comment was really long, borderline necroposting. Reposting seems fine by me, especially if you add the previously-previously-previously links to prior discussions. I just dislike when a fresh discussion gets “merged” into the old story. I like more discussions, never mind the fragmentation. But that’s me.

            I’ll try to ask push (during his next office hours) to determine something like median-delta-of-(comment - post).timestamp or whatever. I can’t imagine many comments are near that 90 day limit. As to why it’s 90: I’ve no idea.

            [1] https://github.com/lobsters/lobsters/blob/b9b477fbbde3706122ef67d94e1d169eadf6f71d/app/models/story.rb#L137C14-L137C24

            1. 1

              What do you dislike about necroposting and merging?

              1. 3

                Re: necroposting, I’m probably just never going to see your comment. It’s basically straight to /dev/null. If it does reappear on /active, I gotta scroll to find the 1 unread.

                Consider that, for the last ~10 comments on any given story, I will probably never see them. If I’ve read thru ~90 comments on Gitsoft acquires Macrohub, I’m simply not going to check out the last ~10. Sorry. I’m not going to rewatch a whole movie for 1 deleted scene. Ain’t nobody got time for that.

                Re: merging, I’m likewise not gonna see it. If your “new” story gets merged into a 2-month-old story, I’ll assume there’s nothing new. When 2 fresh discussions get merged, there’s a lot of duplication (although, that’s the one appropriate merge I don’t mind so much).

                Perhaps I just prefer .. dare I say .. 4chan-style mechanics where newest is at the top & oldest is straight in the trash, once it falls off Page 10. If you wanna “simply” continue a conversation I had 10 years ago, nah I’m good.

                Does that make sense ? I hope I don’t sound like a jerk. I’m just a guy on a MacBook that kinda likes this place

                1. 1

                  I hope I don’t sound like a jerk.

                  Nah, perspective’s important! I hadn’t actually considered discoverability in this sense, focused on my own workflow of searching for opinions after reading (or rather, reading my old notes about) big name papers etc.

                  Anyway, the proposal/idea didn’t awake much interest, so it’s neither here nor there.