Beefy Boxes and Bandwidth Generously Provided by pair Networks
more useful options

The fickleness of Reputation

by ton (Friar)
on Apr 06, 2001 at 05:10 UTC ( [id://70369]=monkdiscuss: print w/replies, xml ) Need Help??

OK, so I'm new to this site, but the randomness of 'Reputation' on nodes has been bothering me, especially as it relates to a monk's experience. Let me describe why:

It is my understanding that a monk's experience level has a high correlation to his/her Perl proficiency (normalized by the amount of time s/he's been on the site, of course). This is a great idea, as it makes it easier for a Perl neophyte to learn which experts are, in fact, experts. The number of experience points a monk has is primarily affected by the reputation of the nodes which s/he has written. A more experienced monk is thus one who has highly reputable contributions to the Perl monks' community. So far, so good.

My problem is the randomness with which nodes seem to be rated. I only have access to my own data, but I believe my results are similar to those other monks have (please let me know otherwise). Let me give two concrete examples:

  1. This node gives a method for computing modular exponentiation on Bit::Vectors. It uses a non-brute force approach that reduces the computation time from O(n) to O(lg n). As of this writing, it has a reputation of 1.
  2. This node is a response to a question on why "if(int(rand))" always returns false. As of this writing, it has a reputation of 9.
IMHO, the former demonstrates a much greater grasp of Perl, mathematics, and overall programming ability, yet I am much more likely to (eventually) gain experience for the latter.

I understand that reputation is meant to be an indication of how useful a node is, not how technically difficult. But reputation is translated into experience, which is meant to be an indication of techincal ability. Or at least that's how I use it, and I think other new users do the same.

I also understand that the reputations I've listed above are subject to change, especially given that the latter node was posted this week. But I see no reason why the trend depicted here would reverse itself over time.

Does anybody else have a problem with reputation of nodes translating into experience, and thereby an indication of Perl expertise? Or am I just being anal?


btw: I haven't thought of a better way of measuring Perl expertise. And this is the best site I've seen at so doing. I just want to make a good thing even better.

Replies are listed 'Best First'.
(jcwren) Re: The fickleness of Reputation
by jcwren (Prior) on Apr 06, 2001 at 05:47 UTC

    Yes, you're being anal.

    But to clarify, there are a number of things that factor in. These are:

    • How well you express yourself. Poorly written nodes are much more likely to be ignored or hammered than even well expressed nodes that are incorrect.
    • Technical accuracy. A node that is technically correct will do better than an incorrect one. However, a good node, deep enough in a thread, usually won't do well because people have gotten tired of the thread. Except for the golfing type nodes, generally the longer a node goes on, the less useful it becomes. Peoples time and attention span have to be factored in.
    • Who wrote it. One might say sad but true, but the fact is there are people that are less respected than others, and usually because they've brought it down on themselves. If your name is DiscoStu, you're doomed. The odds of ever having a positive rep node are somewhere between slim and none.
    • Whim. People are unpredictable fickle creatures. What I may think is an excellent node, someone else will hate. Me, I can't stand 99.4% of all so called 'Perl Poetry'. If it's really bad, I vote it down. If it merely sucks, I ignore it. I think I ++'ed one, once. It was a fluke.
    • Whining/complaining/commenting on reputation of a node. Generally, if you complain about it, people will visit untold --'s on you. Sorry, but it happens. It's worse if you bring it up in the chatterbox, or so I believe.

    Frankly, people assign way too much value to reputation. It's an imperfect system, and always will be, because people are involved. Folks aren't always objective, in spite of the the 'Vote the node, not the person' mantra. $0.79 and 10,000 XP will get you a cup of coffee at McDonalds. XP don't mean jack. It won't buy you a better car, hasn't gotten anyone a better job (yet), and is non-transferrable. If you're attaching your self worth to your article reputation or XP, you need help. Badly.

    Oh yea. Some people, like Dominus, wander in here out of the blue, and accumulate XP like no ones business. That's because they're well known outside of the monastery. Just look in the front of Damian Conways Object Oriented Perl, you'll see thanks to Dominus. Me, my name isn't there. No one knows who the hell I am, and I had earn my disrespect. Pretty much just like everyone else.

    The key to doing that? Good nodes. What's a good node? That's a whole 'nother diatribe, my friend, a whole 'nother diatribe...


    e-mail jcwren
Re (tilly) 1: The fickleness of Reputation
by tilly (Archbishop) on Apr 06, 2001 at 06:34 UTC
    I would say that reputation is closest to being a measure of how much the site as a whole feels that you have helped them.

    People who are competent have an advantage, of course. Being able to tell people useful things is helpful. Being willing to be involved is also good. But a relatively new Perl programmer who is willing to try may gain in reputation more quickly. Indeed I know that several of the saints feel that way about themselves.

    So it would be wrong to assume* that just because I have a higher reputation than Dominus, I am better than him. This is even true when you see that he has been here for longer and I have several times the reputation! It would be accurate to assume that people think I have contributed more overall value to the site. Which is not surprising when I have several times as many posts as he does.

    So what is the difference between competence and what is measured in reputation? Well here are some items:

    1. Cumulative time spent. That is obvious.
    2. Amount of effort put into the site. If you post a lot it will add up.
    3. Early useful answers. If they are early, then you catch a lot of people who only visit the thread once. This often works even if the answer is obvious to an experienced programmer. This effect is so strong that I try not to take advantage of it. I would like to let people with less experience take advantage of these.
    4. Discussions that people like. People are social, they vote up things that they like. This is another aspect that I don't like, but I have to live with. For instance The nature of work makes people laugh, The path to mastery makes them think. Guess which has the higher reputation?
    And so on. Things that will not factor in:
    1. Nodes that appear after the discussion moves on.
    2. Nodes that people do not understand - no matter how good the technical material may be.
    3. Nodes whose tone puts people off - again no matter how good the technical material may be.

    So no, reputation doesn't measure competence. Not by a long shot.

    Nor should it.

    * It would be accurate to assume from the way that I said that that I consider him more competent than me...

(dws)Re: The fickleness of Reputation
by dws (Chancellor) on Apr 06, 2001 at 05:42 UTC
    Does anybody else have a problem with reputation of nodes translating into experience, and thereby an indication of Perl expertise? Or am I just being anal?

    Well, if you put it that way... :)

    Seriously: There are several reasons -- aside from the content of the node -- why a node's reputation isn't a valid reflection on its value:

    1. Nodes have different values to different people. What you find vote-worthy, others might not.
    2. Replies that get posted late in a discussion have fewer eyeballs going over them, and thus have less of a chance of garnering votes. There are some great nodes that were posted weeks or months after a discussion was over.
    3. Replies that happen deep into a discussion are often overlooked by people who haven't taken the trouble to figure out what Some notes below your chosen depth have not been shown here means. (See here for one explanation.)
    4. Once a discussion goes on for a while, and some good answers have been posted, people might just give up on the thread, believing that it's been well mined out. I catch myself losting interest in some threads once they descend into Perl golf.
    5. It may be the case (i.e., here's a hypothesis) that votes from East Coast monks get used up earlier in the day, and aren't available for the many excellent posts made by West Coast monks.
    6. Etc., etc. (See jcwren's post below for more good reasons.)

    The point is, in general, the number of votes that a node gets is not a great reflection of a node's "value". And I don't think anyone is highly motivated to try to make the situation "fair", at least not once they've thought through the implications.

    But life here is more interesting at the edges. Nodes that gather high scores (Best Nodes)or negative scores (Worst Nodes) are worthy of study, not so much for the value of the node (though many of the highly ranked ones are indeed quite good), but as a read on the "pulse" of the monestary, and for tips on which styles of writing and rhetoric people respond to.

      This has been a topic of discussion on another forum that I participate on. We don't have an XP system or a rep system, but it is generally accepted that there are some all time posts.

      Recently, we had a discussion about asynchronous communication (which is of course the style of any online forum) which was the result of two or three really good posts that garnered no discussion. After a week and a half went by, the original poster followed up with a statement to the effect that she was unhappy that the posts hadn't generated the amount of discussion that could be expected of the topic.

      It was noted that the beauty of the forum, and all online forums is that when people are too busy, they won't necessarily just drop the post, but make mental, physical, or digital note of the post (like the Personal Nodelet), and get back to it when they have a) more time or b) something to contirbute.

      A month ago, I kept getting "random" XP, and while I can guess some of it, I know some more was probably related to the upvoting of old posts that provided insight, solution, or possibly another approach to a problem.

      Lexicon commented to me that he gets to vote 6 days worth of votes over his 5 day workweek (see his location to figure out how he accomplishes it). I don't know that the timezone matters too much, because I notice that at the end of the workday, I still have a dozen votes to spare, usually.

      So, while your points are valid in an Mtv generation world, I think that with experience (of using the monastery), old nodes get mined for information and get voted well after the fact. I know some nodes probably do suffer from the Mtv style voting (you mean that node exists??? But it's not on Newest Nodes!), it still can't explain the "random" XP that comes up in the XP nodelet every once in a while.

      ALL HAIL BRAK!!!

Re: The fickleness of Reputation
by geektron (Curate) on Apr 06, 2001 at 05:47 UTC
    you have a valid point re: reputation, but you're leaving one thing out

    who would you rather get advice from -- someone who can blather on about high-level ( or low-level ), obscure mathematical and quantum topics, or someone who has demonstrated a solid working knowledge of the language in question? reputation ( as i see it ) isn't intended as a show-off factor.

    granted, i'd rather have both, but if i need to know why if ( int(rand) ) is always false, i don't want to hear about how i can create an alternate solution that's O(log(n)) rather than the O(n) solution i was working on. . . .

    ( esp. since O notation is completely lost on me. . . :-)

Re: The fickleness of Reputation
by mothra (Hermit) on Apr 06, 2001 at 19:42 UTC
    If you somehow relate XP to Perl expertise, you're way off. :) In fact, I have no understanding of how anyone with more than 3 hours of programming experience would see a relationship there. There are some people on this site (whom I know personally), that do excellent work, and yet have low XP. skeight is a good friend of mine and an amazing hacker. He was the type that would finish his programming exams in 15-20 minutes when everyone else would take 1.5-2 hours (I wish that was just an exaggeration).

    I happen to have reasonably high XP even though at this time I do almost no Perl coding: my day job pays me to do something else and so I focus on that "something else" much more than I do Perl. As it turns out, I am currently doing a website in Perl for a client (it's still in it's very very very early stages, and the admin hasn't configured Apache properly yet, which is making matters worse, but it'll get there). As well, I just love using Perl, I find this site to have some really interesting stuff and think that I can offer useful advice also.

    I wouldn't say you're being anal so much as naive.

Re: The fickleness of Reputation
by coreolyn (Parson) on Apr 06, 2001 at 18:33 UTC

    High expirience is ONLY a reflection of a users commitment to participating in the free exchange of Perl knowlege and values, as it should be. You can't get allot of expirience without posting, and if you post garbage you don't get any reputation. The system seems to work.

    Your main 'block' on this is that you think that expirience/reputation translates to Perl Expertise While there is a wealth of Perl expirience on the site, some of us are truely participating in the site to 'gain' expertise and our expirience relfects how much we keep banging at that process.

(jeffa) Re: The fickleness of Reputation
by jeffa (Bishop) on Apr 06, 2001 at 23:08 UTC
    I am responding a bit late to this one, since everybody has responded in a general manner, I'll be specific and give you my opinion on why those nodes don't have a very high reputation:

    Re: modular exponentiation with arbitrary precision binary numbers (did I say RSA?)
    Correct me if I am wrong, but if neshura editted it (ten days after you posted, from the looks of it), I would have to guess that it was originally posted without code tags. Time goes on, nobody votes it up because it is not properly formatted - by the time it is editted, everybody has already seen it and doesn't think to look at it again.

    Re: if condition only returns false . . .>
    Good, correct answer. But very brief. The only monk I know that consistantly uses brief replies to their advantage is the ubiquitous I0.

    Those are just my assumptions and opinions - I merely wish to help out, not razz you. As for general reasons why good nodes don't get the reputation they deserve, well, it's all about timing and association (don't expect replies to a stupid question to get a lot of ++'s - it happens, just don't expect it). .

    I can tell that you have learned a valuable leason of experience whoring - advertise. It works, just go back and look at how much the reps on your two nodes have already risen. :)


Re: The fickleness of Reputation
by scottstef (Curate) on Apr 09, 2001 at 04:51 UTC
    I chuckled when I read this node. My claim to fame is being the most Perl illiterate monk here. My opinion of perl monks is that it is a community where people can get together and learn/exchange ideas regarding Perl and factors affecting its use. This site does use a reputation/experience point method for kicks. Personally, I don't think that much of the whole rep issue, although i don't know how I would act if i ever got a lot of -- votes. I do look in the chatter box at people ranked above/under me and look at their home nodes.

    Some of the members here merlyn, davorg and others here that I am forgetting, have written many publications that I have read and learned a lot from. Others have written stuff on their home nodes that intrigues me, make me look for stuff they post here.
    As far as the node you posted, regarding finding a very eloquent solution to a math problem, I really don't have much background in math, don't think I will ever need such a solution, so I apologize for glancing over it. When I look for nodes to vote on, I look for an interesting subject, does the node appear to have something in it that interests me, something I may need to hack together one day? If so, I will probably read it. Unfortunately I wish I had the time to read every posting here, realistically, I may get an hour or so daily to do so. Don't take a low node as personal affront, it probably means you have written above the community's head, in theory if I understood your code, I would probably would have voted on it. If you had been down voted, then yes, I would see your point.

Log In?

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: monkdiscuss [id://70369]
Approved by root
and the web crawler heard nothing...

How do I use this?Last hourOther CB clients
Other Users?
Others meditating upon the Monastery: (4)
As of 2024-04-21 12:15 GMT
Find Nodes?
    Voting Booth?

    No recent polls found