To help not to misguide

[c_chipster]

This is not meant to be a popular node, but I beg your people to keep this post, and allowing opposition as long as it bear truth.

Very often, people post answers that are wrong, I hate that. Although this is not altogether avoidable, because of different levels of tech background, but at least be nice not to answer questions that one is not familiar with.

When some of us came for XP points, most only came for question and answer. Please leave those people alone, and don't play XP games on them. There are plenty of opportunity for XP point, but restrain yourself to only provide correct answers, even duplicates would be fine, at least confirms the correct answer, and at the same time gain some XP points. Harmless at least.

[revdiablo]

If you see an answer that is wrong, reply to it and explain why it is wrong. That's how it works around here.

I think you are assuming far too much with your premise that wrong posts are about XP. Often times, people think they know the answer and are wrong. Or they take a guess that turns out to be untrue. Or they misread the question. Or many other things... So please adjust your assumptions.

[bobf]

Very often, people post answers that are wrong, I hate that. Although this is not altogether avoidable, because of different levels of tech background, btu at least be nice not to answer questions that one is not familiar with.


You are correct: it is not uncommon for a thread to contain a reply that is not necessarily 100% correct. There are many reasons for this, and your statement highlighted one of them. In addition, the author may have misinterpreted the question or posted an answer after the OP updated the question. In general, replies are offered with the hope of being helpful to the OP. This site provides a place for people at any position along the Perl learning curve to ask questions and provide answers.

We encourage people to ask questions (after reading the appropriate documentation and using Super Search, of course), since PerlMonks would not be what it is without them. Similarly, we should encourage people to post answers. IMO, the richness of the site is derived in large part from the diversity of responses. Some replies may offer a snippet of code, some may suggest or explain an algorithm, some may provide a technical explanation that provides a glimmer into the guts of perl, some may provide links to previous threads or off-site resources, and some may touch on other areas. Responses are based on the authors' experiences and knowledge, and they do not all approach the same problem in the same way. Individually, each reply might not answer the question completely, but hopefully all replies taken together will address all of the important parts of the question.

IMO, even partially-correct replies have value. Such replies typically encourage a post from someone else, who gently corrects the errant information or points out a caveat that might have been overlooked. Occasionally such a post will initiate a spirited discussion that delves into a dusty corner of the language or uncovers a new quirk. Most of the time those subthreads go well beyond my understanding, but I still find them useful because it gives me another opportunity to learn. PerlMonks is about learning: by reading threads, asking questions, offering solutions, and every now and then by being corrected.

As an aside, and prompted by your reference to the "ugly thread about Oracle" in Re^2: To help not to misguide:

IMO, it is not the initial "incorrect" reply that makes that thread ugly, rather it is the insulting posts by Anonymous Monk that followed. There is a difference between correcting an errant statement and launching a personal attack on the author. The former helps to educate everyone that might have had the same (incorrect) thought as the author, but the latter only clutters up the thread with posts that contain no meaningful content.

[GrandFather]

There are a slew of reasons that people post replies that are incorrect in some part. Most often it is an oversight and almost always the problem is picked up and discussed by others.

Imploring people to only provide "correct" answers not only doesn't fix the problem, it actually lowers the quality and value of PerlMonks! One very important aspect of the PerlMonks community is that discussion of questions takes place on many levels and therefore matches well to the many levels of understanding of the various monks. A simplistic answer may miss subtle but important issues, and therefore be "wrong", but will very likely provide a stepping stone to greater understanding by the author of the "wrong" node as well as greater understanding by other monks at a similar level in their journey to mastering Perl and its environs. Wrong answers are often much more valuable for this reason than a simple reply giving some correct code without exposition.

Reputation is important here. It encourages people to post and helps maintain high quality posts. The XP system expands in node reputation by providing a sort of global reputation based on participation and quality of participation. The system is somewhat open to abuse, but generally works very well. A major reason that the monastery is such a welcoming and helpful place is that the XP system encourages that atmosphere and encourages participation.

Remember too that a node's reputation is not available until after you vote for the node except through best nodes. By the time a reply accrues enough ++ votes to hit best nodes it has been seen by a good number of people and it is pretty unlikely that a significant error will go unnoticed.

The system currently works well. Wrong answers seldom go unanswered. Leave well enough alone I say.

[blue_cowdawg]

Very often, people post answers that are wrong, I hate that.

Very often? What is your definition of very often?

In my experience the number of right answers I've seen to questions far outwieght the number of wrong answers I've seen. If you were going on about wrong questions I'd be right there with you.

The few wrong answers I've seen have have been dealt with via the experience system or by follow up answers contradicting them. There certainly are monks here that won't hesitate to point out when someone is wrong. Some more harshly than others.

I'd also rather see a wrong answer to a question than no answer at all. Although if I don't know the answer to something I'm not liable to post a follow up if I am not sure of my answer. Still a wrong answer can be an opportunity to learn.

How can you learn from a wrong answer? Well... first off if you know enough to know that an answer is wrong then follow up yourself and point out what is wrong with the answer and post corrected information! This provides us all an opportunity to learn something.

Lastly, you never know your own perceptions of an answer might be wrong. You follow up with a post and say "NO! Grass is orange!" and you can be sure you are going to find out in glorious detail that no, grass is green. Unless it hasn't been watered in a while... and then it's brown. :-)

Anyway... don't get too bent out of shape if you see a wrong answer. Post a correction and let the exchange of information, knowledge and ideas happen. That's what makes PM so great!

---

[jZed]

>> Very often, people post answers that are wrong

Perhaps you should go to www.everyonehereisperfect.com. In the real world, people make mistakes. Learning is not a process of always getting the right answer, it is a process of gathering information about a subject from right answers, semi-right answers, semi-wrong answers, wrong-answers and everything in between. Sometimes a wrong answer makes an assumption about the problem and just seeing that assumption sheds light on the problem so the wrong answer ends up being helpful. If you can't learn from mistakes (your own and others'), you can't really learn.

[g0n]

Wrong answers can add value because:
• The person who posts the wrong answer will almost always be corrected by someone who knows better, and hence will learn from it
• People who thought the wrong answer was right, and just another way of approaching the problem, will see the correction and realise that it was wrong. They also will learn from it.
• Sometimes an original post will not make the question clear. Occasionally an answer will be posted that others disagree with, and the OP will later state that the 'wrong' answer correctly interpreted the question.
• Not infrequently a wrong answer will cause the OP to clarify their original post, leading to better answers.
• People who post rude and personal responses to a 'wrong' reply will often get downvoted, and perhaps learn to deal more gently with people.

This is not a commercial support site, its a community forum. Everyone has something to contribute, everyone has something to learn.

Update: inserted the bold word 'can', in response to PerlMouse.

Just to clarify, if the OP gets 2 right answers, 2 wrong answers, and those wrong answers get highlighted, then at least 3 people have learned something. How is that bad?

[spiritway]

I can't speak for anyone else, but when I answer a question, I answer to the best of my ability, with the knowledge I have today. If we all waited until we knew Perl perfectly, none of us could answer. Of course our knowledge is imperfect. It always will be.

The problem is that it's very difficult to know what you don't know - to realize that your knowledge about something is faulty.

The XP system is imperfect, of course. Sometimes mistakes get upvoted, often because the answer is a common mistake that many people make. It happens, and sometimes you get complete nonsense passing as solid gold.

It seems to me, though, that more often someone out there knows what s/he's talking about, and offers a correct answer, or a correction to a mistaken answer. Then you have to decide who to believe - and the right answer might not get as much XP for any of a number of reasons.

Bottom line: at the Monastery, as in real life, we have to use our own minds, come to our own conclusions, do our own reality testing. Relying on XP to guide us is misguided, because XP is mostly a measure of a post's (or poster's) popularity, not its accuracy.

Finally, after a few erroneous answers, you get to know who has a clue, and who's just guessing. Even the experts make mistakes, though, so you still have to check everything yourself.

[holli]

This is not meant to be a popular node, but I beg your people to keep this post, and allowing opposition as long as it bear truth.

This struck me a little. I don't know a place where the power of deleting a post is handled as responsible as here. Even a clear troll node (here in this thread) that says "Fuck you all" in bold big letters gets only "reaped", meaning that one can still access and read it by one more click. In theory, nodes can be totally deleted (nuked), but I haven't seen that yet and probably will never.

This is not XYZ-forum with a single admin.

[demerphq]

In theory, nodes can be totally deleted (nuked), but I haven't seen that yet and probably will never.

Site policy is that we don't nuke nodes. Even duplicate patches don't get nuked despite the fact it would make no difference to anybody if they were. In the olden days we used to nuke occasionally, but these days the policy is that once a node is created it can change type but will never be destroyed.

[BrowserUk]

Perhaps we could tackle this "problem" from a different direction. Head them off at the pass so to speak, by limiting the possibility of erroneous replies.

Add a new button to the consideration nodelet that is only presented to those monks that have never posted an incorrect answer.

All new nodes would be locked, and no one would be able to post a reply until that select band of Uber-monks have "signed the post off", by clicking the above button, indicating that they either do not know the answer, or cannot be bothered to post one.

Only once all Uber-monks have so signed a post off, would that post become unlocked. and so become open to the rest of us mere mortals to post our thoughts, half-remembered plagerisms and wild-arsed guesses.

I was half way into coming up with an elaborate scheme for determining who should be afforded the privaledge of Uber-monk; by analysing their XP/posts ratios, probably limited to first child replies or similar; a cross-authorisation scheme; and a few other frills; when it struck me.

Most of them seem to already know who they are, so maybe we should just let them "sign up" for Uber-monk status via a button on their preferences?

[TrekNoid]

Sorry for jumping in late, but I've been away for a while... just saw this today.

Doesn't your request assume two things?

1.) First, that there's some universal definition of 'correct' that we all can agree on?

There are plenty of 'correct' answers that don't work in some situations, and there are occasionally 'wrong' answers that work fine...

Some people believe that 'correct' means it has to work under 'use strict', others don't... some believe that CPAN is your best friend, and use it a lot, some like to roll their own...

'Correct' is in the eyes of the philosopher... TIMTOWTDI is in the eyes of the Perl developer :)

2.) Aren't you also assuming that we come here only for answers, and not necessarily to learn?

I posted an answer once, using $#ARGV to check for number of arguments... It's just the way I learned to do it... It *worked*, and wasn't incorrect, but merlyn saw it and quickly pointed out that @ARGV was a much better way of doing what I suggested.

So guess what? Not only did the OP get a better answer, *I* learned something too.

So yeah, I posted an answer that would be considered 'wrong' by most standards... but the OP, myself, and anyone else who happens to view that thread later get to learn more than just the answer to the original question.

Trek

[kulls]

"Very often,people post answers that are wrong"

[eric256]

Hey Kull,

It appears that when you are replying to nodes you are putting the message inside the HTML tag labeled <div class="pmsig"><div class="pmsig-263384">. Many of use have our settings set such that the signature (the code inside that tag) is not shown, or shown small or lightly. This means that 95% of your message is unreadable to a fair amount of us. I wouldn't normaly have replied like this, but i've messaged you several times with no response. If you are new to HTML and this confuses you please check out the chatter box or fullpage chat to get some assistance. I just don't want your messages to go unread because of such a silly problem.

[kulls]

hi,
i'm really sorry for the inconvenience and thank you for your suggestion.but
I wouldn't normaly have replied like this, but i've messaged you several times with no response.
I guess i didn't get any messages from you.
- kulls

[TedPride]

 ?? ??Dt downvoted (and sometimes flamed). This is enough to keep most people from posting them, or at least from posting them the next time someone asks the same question. The real problem is not wrong answers anyhow, but rather answers that say nothing, or that answer a different question than the one originally posted. I've occasionally done the latter by mistake myself.

[Anonymous Monk]

levels of tech background, but at least be nice not to answer questions that one is not familiar with.

Why? To avoid potentially giving a wrong answer? Wrong answers are given all the time even when you're completely familiar questions.

[phaylon]

Ack. I think the better solution would be to ask people to state if they're not really sure. But then, I often see 'IIRC's, 'AFAIK's, '(untested)'s and other marks that tell the questioner if this is a hint, a suggestion, a solution or whatever. In fact I think PM is quite good at self-regulating.

[Anonymous Monk]

I agree with upstairs, there are too different problems here: bad answers and XP points. Although in a way, it is sad that bad answers sometime gain XP points, and mislead people to believe that was correct after all.

Considered (planetscape): SSDD, anonytroll
Unconsidered (holli): Enough Keep votes (Keep: 9, Edit: 0, Reap: 10)

[sanPerl]

I don't agree that mis-leading answers should be discouraged.
The fact is, many times monks at different levels give different answers (and they are very honest), they think that they are right. Ultimately some guru provides the correct answer and the question ends. This is how any discussion goes in real-life also.
We should encourage monks to talk, discuss and to be BOLD. Even Wikipedia (a Free web-encyclopedia) asks its members to ‘Be Bold’, it allows wrong information (There are many reviewers who correct the wrong information).
Also I believe that Monks who posts questions are wise enough to test the solutions, before implementing it in their projects