Support » Requests and Feedback » RFC :: New Feature Query :: FAIR Comment Ordering

  • Summary
    Blogs with heavy comment traffic often include a number of comments whose insights may rival or exceed that of the blogger’s own. Such comments represent valuable content worth reading but can get lost in the voluminous comments of well-known bloggers. Even if the blogging system includes “liking” for comments, the ordering of the comments themselves — whether sorted newest-to-oldest, oldest-to-newest, or even by “liking” count or some other metric, results in an uneven probability skewing for the *order* in which the comments are presented and read, and hence any “liking” metric derived from it.

    Obtaining a fair ranking of the comments themselves (not the individual blog posts) would be of great value, because it would allow favored comments to rise to notable levels via standard ranking metrics. This fair ranking could be obtained using probability distribution mathematics that rearrange the comments in a “fair reading order” that takes into account a number of dynamic variables which are nevertheless known when a reader chooses to read a blog post: total number of comments, position of each comment in the order, probability distribution curve of how many comments might be read by a current user. (This system assumes only one level deep comments). Using these variables in addition to The algorithm would arrange the comments for reading order to a reader such that over the lifetime of the blog post, the probability that each comment would be read would be equal. This would allow for “fair voice”, overcoming the current problem that the first (or last) to post comments receive a disproportionate amount of “reads.”

    Request For Comments
    If such a system were to be integrated into WordPress, would it constitute changes to the core, or could it be done as a plugin?

Viewing 4 replies - 1 through 4 (of 4 total)
  • Moderator Ipstenu (Mika Epstein)


    🏳️‍🌈 Halfelf Rogue & Plugin Review Team Rep

    This can (and has been) done as a plugin. Many plugins can do this.

    Comment Rating
    GD Star Rating

    I searched for almost an hour and could not find a single plug-in that does what I suggest. I could not find even a simple plug-in that does an elementary form of what I suggest. I have never seen what I suggest anywhere on the web. Can you show me either of the following:

    (1) A web page that does what I suggest?
    (2) A WordPress plugin (you say there are many. Name one.)


    Moderator Ipstenu (Mika Epstein)


    🏳️‍🌈 Halfelf Rogue & Plugin Review Team Rep

    Sorry, I mis-read what you want. The plugins I listed all do ranking.

    But. There’s nothing that does ‘fair comment ordering’ but that would be because there’s no equation that woud work for all sites in order to determine that. What YOU would say is fair order I may not and so on and so forth (even if you make the comments same-level, people still want to reply and have a discussion). Way too many variables, plus many comments are replies and they would have to be weighted.


    Thanks for reading my post. If we begin with the assumption that the order in which people originally post their messages should have no bearing on the order in which such messages are presented to future readers (ie: first voice is not more important than last voice, which is reasonable), and also assume there is some way to gather the data of how many messages each poster may read (via click-tracking), then there is a deterministic solution which can proscribe how each set of messages should be displayed to any particular poster who wishes to read messages, such that over time, all messages will receive an equal probability of being read. The solution can be derived using a discrete approximation (and limit theory) of the non-linear probability curves of:

    (1) posts measured against time
    (2) number of messages read by readers measured against time

    Proof that such a solution exists is not too difficult. For instance, if you post the messages oldest first, it’s clear then that message reading (and any “liked” metric derived from it) is slanted towards the oldest posters. So, blogs invented the “newest first” to take away the advantage of early posters. Yet this isn’t completely fair either, because once the number of messages reaches into the dozens and beyond (exceeding the number of messages any single poster is likely to read) the messages in the middle of the pack aren’t as likely to be read. So whether you read from earliest to latest, or vice versa, this skews the reading of messages toward the start and end. The characterization of these posting and reading patterns is a non-linear probability distribution as a function of time, and can be discretely characterized using non-linear bin-sliced probability distribution functions (pdfs). These curves can be calculated by extracting the frequency of posts for a particular article in conjunction with message reading frequency. Or, in the case of different pdf, the set of characteristic reading patterns can be kept and curve fitting used to dynamically build a correspondence between a live blog post (with messages) and the likely final pdf.

    Since these two pdfs are likely to be non-uniform across web sites (and even authors, days of week, etc.) I would agree with you that technically there is no “one” equation. However, the pdfs of each article, and web sites, are real numbers that can be computed dynamically using standard probability theory.

    Minimally, a much fairer way with less advanced mathematics would be to simply randomize the list of messages after receipt of each message, and skew towards the “most recent received” by the inverse of the reader rate PDF.

    I’ll take it from your post that you don’t know of any plug ins that currently do this.

    My next task is to work out the math and the approximation formuli.


Viewing 4 replies - 1 through 4 (of 4 total)
  • The topic ‘RFC :: New Feature Query :: FAIR Comment Ordering’ is closed to new replies.