in reply to Re: Choose a maximum for Rep
in thread Choose a maximum for Rep

The reason I came up with this is because I find some nodes worthwhile, just not that worthwhile.

I disagree that votes become more or less valuable. They're as valuable as you think they are. Another option is if you vote ++<10 and it's already there, you don't spend a vote.
Several monks actually -- votes that are at certain levels because they feel that the rep has exceeded the actual worth.
Sometimes you want to vote on a node just to see the rep and this would allow that.
Of course you could extend the logic so that a vote with a cap of 10 would -- a node that already has 20, or set up some mathamatical expresion if you want to get more complicated.

Anyway, it's just an idea. I'm not attached to it.

Replies are listed 'Best First'.
Re: Re: Choose a maximum for Rep
by chipmunk (Parson) on May 25, 2001 at 07:28 UTC
    I believe that, under your proposal, votes cast later would be more valuable than votes cast earlier.

    Here's a statistical proof. This code tries random orders of voting, keeping track of where in the order your vote is cast, and calculates the average reputations.

    #!/usr/local/bin/perl -w use strict; $| = 1; my @all = ((5) x 6, (10) x 6); # twelve voters: six voting up to 5, six voting up to 10 for my $own (5, 10) { # your own vote my $i = 0; my @others = grep $_ != $own || $i++, @all; # remove your vote from the list for now... my(%r, %n); for (1 .. 5000) { my @voters = shuffle(@others); # shuffle the other votes splice @voters, my $pos = int(rand @voters+1), 0, $own; # put your vote back in, and save the position my $rep = 0; for my $v (@voters) { $rep++ if $rep < $v; } # calculate the reputation $r{$pos} += $rep; $n{$pos}++; # adjust the stats for this position } print "$own:\n"; for (sort { $a <=> $b } keys %n) { printf " %2d %4d %4.2f\n", $_, $n{$_}, $r{$_} / $n{$_}; } # print position, number of occurences, and average reputation } sub shuffle { my @list = @_; for (my $i = $#list; $i >= 0; --$i) { my $j = int rand $i + 1; @list[$i, $j] = @list[$j, $i]; } @list; }
    And some results (your position / occurences / mean reputation):
    5: 0 411 8.77 1 404 8.80 2 409 8.77 3 414 8.74 4 411 8.76 5 452 8.25 6 393 8.36 7 425 8.28 8 409 8.32 9 417 8.31 10 444 8.29 11 411 8.23 10: 0 386 8.17 1 439 8.15 2 420 8.25 3 435 8.14 4 393 8.21 5 407 8.74 6 437 8.73 7 412 8.71 8 444 8.69 9 402 8.67 10 399 8.77 11 426 8.75
    Observe that the node's reputation will be closer to what you personally want if you vote later. If you're voting low, a later vote leads to a lower rep; if you're voting high, a later vote leads to a higher rep.
Re: Re: Choose a maximum for Rep
by footpad (Abbot) on May 25, 2001 at 20:49 UTC

    Several monks actually -- votes that are at certain levels because they feel that the rep has exceeded the actual worth.

    If this is true, I do not see how this is any different from the personality voting that has been railed against so strenuously. Instead of voting (or ignoring) a node's content, you are voting its Reputation.

    If those doing this are against personality voting or would rather see more votes given to code, then I ask them to reconsider this practice. I do not believe it's necessary to "take a monk down a peg" for anything other than a bad node.

    Vote the node's content, not the poster nor the node's reputation.


      I do this quite a bit. I don't understand how this is "taking a monk down a peg"? The monk got 64 votes and a few dozen XP for a too-short, too-simplistic, mostly-but-not-completely-correct reply. After all of that, my downvote doesn't hurt much at all.

      Better than that, it helps (in some tiny, tiny way -- which is about all you can hope for with a single vote) to discourage monks from posting too quickly "just because" short, easy-to-understand, fast replies tend to get more ++ votes due to community dynamics.

      I also sort replies by rep and so I sometime downvote nodes that are "listed too high" in comparison to other replies.

      And none of this is "voting the reputation". I vote the node contents but I take the node's reputation (and the node's author) into consideration!

      I'm very reluctant to downvote newbies as that just leads to a bad initial experience when they don't even know enough to learn from the downvote. I'm relatively reluctant to upvote nodes by those with high XP (I know they can post "good" nodes quite a bit, so I want to see something that is "a good node, considering who wrote it") or that have recently posted a ton.

      I upvote nodes that have too low of a rep when I think I understand the reason for the low rep and that I don't agree with that reason (for example, if a node dares to say something negative about "perl" but is accurate). In a "discussion" I will only upvote the very best of the thread while in "code" threads I may upvote the entire thread because I find that "discussion"s get higher reps than "code" threads (if that changes, then so will my voting).

      So I'm left to assume that footpad does a sort of mindless, feel-good upvoting only and doesn't take much of anything complex into consideration {duck}. If I had more choices in voting besides -, +, 0, then I might not need to use downvoting as often in order to satisfy my desire to have my votes make sense to me.

      I do make an effort to not downvote in a way that is likely to result in an "Ack!" message nor in a node with a negative reputation, because I don't particularly like either of those except in extreme cases. But, in the current environment of 98% upvotes, I do downvote quite a bit. Just because I vote this way doesn't mean I want everyone to vote this way (if everyone else did vote this way, I certainly wouldn't anymore!).

              - tye (but my friends call me "Tye")

        I suspect we'll have to agree to disagree on this, for I see little difference between the personality voting that's so hated and this practice.

        I do understand your logic; I simply disagree with it. I don't think it's fair to the community as a whole, for your criteria involves something other than the node's quality. As chipmunk said earlier in the thread, "If you don't think a node's worthwhile then don't vote on it." This is part of what I was trying to get it when I suggested people vote responsibly.

        And, for the record, your assumption is slightly wrong. (No need to duck; informed criticism is always welcome.) I do vote nodes down...when I feel the content is wrong, misguided, non-contributing, trollish, and so on. I do think certain nodes are overrated, but would rather (if enough of us are that upset over it) use a different approach to handle it.

        I reward nodes I like, learn from, or feel have generated interesting conversation. But I do not punish nodes because they've been successful. I do not vote them down because I feel the community over-reacted. I simply don't feel that's fair to the poster.

        It's a different point of view. That's all. I'm not saying you're right or wrong. I'm simply disagreeing with the practice.

        If you feel voting a node for anything other than its content is wrong when we call it "personality voting," then how is "fitting a node" to its "appropriate" reputation any different? You're judging it against something other than its content. It's the same thing under a different name.

        I may disagree with the community's overall assessment, but I'm willing to accept the fact that the tribe has spoken. Besides, there are enough nodes that teach me something that I don't really have votes remaining to "correct" their reputations.