Historic:FeatureList: Difference between revisions

From SoylentNews
Jump to navigation Jump to search
(moderation/threadjacking)
No edit summary
Line 35: Line 35:
**# Profit!
**# Profit!
***Actually, not just the immediate parent like I said, but 'all' ancestor comments. Otherwise as step 1.5 I just reply with an AC troll (which someone else will rightfully downmod), then in step 2 I reply to ''that''; now that the original AC first post is my comment's gp, so I can still downmod it.
***Actually, not just the immediate parent like I said, but 'all' ancestor comments. Otherwise as step 1.5 I just reply with an AC troll (which someone else will rightfully downmod), then in step 2 I reply to ''that''; now that the original AC first post is my comment's gp, so I can still downmod it.
** ''' Good point ''' you should be prevented from moderating any children ''''and''' ancestors of your posts (but sibling chains are still fair game)


* Moderation Engine
* Moderation Engine

Revision as of 16:12, 7 February 2014

  • UTF-8 clean for anything that doesn't end up in a URL (i.e. tags, titles, channel names, user names)
    • Of course, some combining marks would have to be filtered, and the resulting text round-tripped through NFD->NFC to prevent certain types of attacks against users or making text difficult to index.
  • Channels (like yro, politics, apple) as first class objects
  • Submission queues for all channels with membership at the discretion of the channel owner; being able to submit also implies up/downvoting submissions and tagging
    • Editors/channel owners only
    • List of users who can submit
    • All of my friends (for user journals)
    • All users with mod status
    • All registered users
    • Everyone
  • Articles are also first class objects
    • Perma-link independant of channel
    • Articles can be cross-posted to other channels by a channel owner (built into the UI if the editor manages more than one channel)
  • Comment key features
    • Comments are first class objects, just as before, just like articles
    • Comments can be edited for a short period of time by owner
      • HOWEVER: Edited comments get a new ID and it links to older versions in the new one. This detail is made prominent to viewer if a comment reply happened before an edit.
    • Allow alternate markup options (bbcode, markdown, wiki markup)
      • Provide a mapping to actual div and styles that will be applied, or HTML equivalent codes
  • Allow moderation in same article that you've posted in; only disallow moderation to your own reply chain.
    • You're obviously not allowed to moderate in your own accepted or posted article (treating all comments as replies)
    • Does "reply chain" include parent? If not, how do we deal with threadjacking:
      1. AC posts first comment on a new article (let's presume it's an upmod-worthy comment, not fristpsot)
      2. I post my unrelated comment as a reply to AC's FP, in order to achieve greater visibility
      3. I downmod the AC to -1
      4. Now AC is invisible to anyone with threshold != -1 (including many with mod points and threshold=0, who would have modded AC up on his own merits)
      5. Unless/until users with threshold=-1 and mod points happen by to rectify things, my comment is now the first thing most users see.
      6. ??? (sorry, can't resist...)
      7. Profit!
      • Actually, not just the immediate parent like I said, but 'all' ancestor comments. Otherwise as step 1.5 I just reply with an AC troll (which someone else will rightfully downmod), then in step 2 I reply to that; now that the original AC first post is my comment's gp, so I can still downmod it.
    • Good point you should be prevented from moderating any children 'and ancestors of your posts (but sibling chains are still fair game)
  • Moderation Engine
    • Moderation tags only (default scores)
      • Overrated -1
      • Underrated +1
      • Offtopic -1
      • Interesting +1
      • Troll -1
      • Insightful +1
      • Flamebait -0.5
      • Funny +0.5
    • Users assign own weights to tags in the range -2.0 -> 2.0 with 0.5 increments
    • System rounds x.5 towards 0 in comment spill / threshold logic, display capped at -1 and 5.
    • A score for the default weights is saved for the comment for use by article spill (for googlebot or "load all comments" from anonmyous user)
  • Moderation strategy
    • Chops instead of Karma
    • Chops are derived from:
      • Articles accepted for submission
        • But NOT articles self-authored on a channel you're an editor of
      • Comments that are replied to by others without a troll modifier
      • Positive moderation (as judged by the mod action with respect to their own point value weights... so if they think funny is bad, it counts against you)
      • Positive metamoderation outcome (see below)
    • Registered users that have used the site > [threshold] days, posted > [threshold] times, and have positive chops get moderation duty
      • The more chops, the more mod points per day, with a total cap for unused
    • Meta-moderation is available to users that have used the site > [big threshold] days and have > [threshold] chops
      • Metamoderation is not "special", a meta-mod capable user can see a random selection of recent mods at any time and metamod.
      • Metamod takes 2 mod points
      • Metamod can spend a mod point to "re-roll" and see a new set of random moderations
      • A moderation is undone when it's "score" goes is negative. It is "reapplied" when it goes positive. If the score dips to -3, the moderation is removed entirely.
      • A user is not rewarded or punished for the metamod specifically.
        • A running total of positive and negative meta-moderations against them is calculated
        • Certain threshold for positive and negative meta-mod counts result in fixed deltas in chops
          • Having > 5 positive metamod could be +1 chop score. >10 +2, >20 +3; > 3 negative is -1, > 5 is -2, > 10 is -3