Thread topicallity, accuracy and integrity....

[TheCPUWizard]

It was recommended that I post this question here to solicit feedback from the community....

Many times a thread will be started with a specific question [if wer are so lucky that the initial posting is specific and provides the necessary information].

During the replies a number of conditions may occur:

1) Information not directly related to the original question is posted (either by the original poster or by someone else)

2) Information that does not conform to accepted "best practices" may be posted which, while meeting the the OP's requirement is "not how it should be done"

3) Information which is downright incorrect is posted (this is probably the rarest of the situations under consideration.

The question is:

"Given one of the three cases above have already occured on a thread, how should members reply?

Going into details that are not germaine to the OP's question can easily be distracting or overwhelming, but what about the impact on the thousands of people who will search CodeGuru over the coming years and find this "less than ideal" information???

Where is the "Best Value" obtained in both adressing the OP's current needs and the benefit of the community at large over an extended period of time???

Looking forward to hearing peoples thoughts....

[mine should be pretty obvious from my posting habits, so I intend to be a "silent observer" for a while...

[kirants]

Going into details that are not germaine to the OP's question can easily be distracting or overwhelming, but what about the impact on the thousands of people who will search CodeGuru over the coming years and find this "less than ideal" information???

Hm.. the first thought that came to mind, given the current situation is, do people bother to search prior to asking at all?

Anyways, assuming people do search and bump into threads that contain a lot of all sorts of information, am not sure how it can be resolved. Often, it is easy to point out the dead wrong ones and it is our responsibility to do so with an explanation of why ( without the why, it becomes just another opinion ) . But the other 2 types can be hard to pin point. Who really decides if the post is slightly off topic. It needs someone with infinite knowledge to be able to reliably say so. The problem becomes even worse because sometimes, even if people did have a point and posted something that appears to be off-topic, won't come back to clarify if someone requests for one. Maybe they did have a point, albeit, something beyond the rest of the visitors.

I personally feel that when something is found to be wrong or misleading, it has to be corrected. It is a responsibility. However, it also helps if one explains why. It also helps if someone is held wrong, they sportively admit it if someone pointed it was wrong convincingly.

[MikeAThon]

I tend to agree with Kiran. If something is downright wrong, it's a responsibility to correct it. Everything else, however, is probably more like an opinion. Best practices are more like "rules of thumb", and thus can be broken for any reason or no reason at all. They are thus a matter of opinion. Likewise, whether a reply is truly off-topic also seems to be a matter of opinion.

And of course you know the old joke about the part of the anatomy that most closely resembles an opinion: "... everyone has one and they all stink..."

PS: I also agree that only very few people tend to search before they post.

[GremlinSA]

It was recommended that I post this question here to solicit feedback from the community....

Okay here are my opinions to each of the circumstances...

1) Information not directly related to the original question is posted (either by the original poster or by someone else)

This is probably the most common 'problem' .. I've seen threads go way off the posted question with the OP using that one thread to solve a dozen problems on his project, to threads where someone else post a seemingly related question on the same thread..

The general consenses that i've seen is to politely ask the person to post a new thread, or else often a mod comes in and splits the thread..

My opinion is that of.. "One thread, One question, Many solutions." There are some fantastic threads where the disccution is directly related to the problem with, "First try this.. didn't work.. ok Try this then ... etc.." These threads are essentially what you want to be in search results..

2) Information that does not conform to accepted "best practices" may be posted which, while meeting the the OP's requirement is "not how it should be done"

Here is the conundrum.. the solution often depends on the OP's level of knowledge ..
One example - Opening a binary data file ... depending of the op's level.

Basic skills. - Use "OPEN Filename FOR BINARY AS #1"

Moderate skills - Use FSO or API's ...

There are three solutions to open and read datafiles..

However, I do understand what you mean by Best practice , and it does not mean how you (or anyone) would specifically do it, it's more of a general way of performing tasks. (oops my mind is blank, cant think right)

I have seen some solutions that sometimes even a beginner will think, "What the #$@% !".

3) Information which is downright incorrect is posted (this is probably the rarest of the situations under consideration.

This is the worst, and i know that i'm often guilty of posting incorrect info, mostly because of speed reading the question, and not getting all the info provided right, or because of confustion between two or more languages..

However, if noticed i will either edit and repair the post, or simply admit to my mistake and accept the consequences..

Going into details that are not germaine to the OP's question can easily be distracting or overwhelming, but what about the impact on the thousands of people who will search CodeGuru over the coming years and find this "less than ideal" information???

Where is the "Best Value" obtained in both adressing the OP's current needs and the benefit of the community at large over an extended period of time???

The best that we could do is to positively (4's or 5's) thread rate thread's who's contents are "Ideal" and also add tags to better the search results. and negatively (1's) rate less than ideal threads.


Gremmy..

[Arjay]

Here are some thoughts on the subject:

To me, it really boils down to answering the questions in the simplest manner.

Not that I always do this, but I believe it's important to judge the abilities of the OP in answering questions.

Sometimes the OP isn't looking for the 'best' or 'best practice' kind of answer because they may not have the skill to understand it on that level. A simple example that comes to mind is when a question is asked about using the XmlDocument in C#. Usually a cleaner, best practice approach is to use XmlSerialization, but again it depends on the level of the OP. I know I have spent quite a bit of time showing one particular OP how to perform the serialization, included modified source code, only to have the OP post a new thread about 2 weeks later that made me understand he didn't get it. In the second thread, I created yet another sample app and remember thinking it's not really worth it to spend so much time.

I'm beginning to think that including sample apps as I have done in the past may not be worth it. I look at the download count for the attachment and have noticed that I don't get too many downloads (< 10) for an app that I've spent 2 hours coding up. I spend all that time and it doesn't really seem worth it.

Folks don't really seem to give points too often either. As far as points, I get most of my points of late from George2 when he's around. It's kind of ironic because I usually torture George a bit when he asks the inane questions (but he still gives me points :-) ).

In keeping with the simplest answering approach, I think it's important not to let the original question get dragged off topic with separate detailed discussions. Years ago Sid and I used to do this (probably more my fault than Sid's). Sure on a purely technical sense we both may have been correct (or the differences certainly weren't worth arguing about), but I doubt our discussions were any real benefit to the original question(s).

I imagine it must be difficult for a newbe to have to wade through 3 pages of a 'side' discussion trying to understand enough to pick out answers to his/her question. I believe very few people benefit from this - maybe the top folks on this site, but they probably don't need to read about it anyway.

[kirants]

I'm beginning to think that including sample apps as I have done in the past may not be worth it. I look at the download count for the attachment and have noticed that I don't get too many downloads (< 10) for an app that I've spent 2 hours coding up. I spend all that time and it doesn't really seem worth it.

IMO, don't rush the judgement. If you felt that it was useful to attach it. So be it. Don't look at the download count. You'll be amazed how many times I have got PMs, emails from people just saying "BTW, thanks for posting that information in that thread. It saved my day" and am sure there are many that didn't specifically email. Just feels good to know that somewhere someone finds some use for it.

but I doubt our discussions were any real benefit to the original question(s).

Hard to say if they are useless. Maybe not for the OP, but someone who bumps into the post later on could find the discussion useful. So, if iit's not insanely off-topic, like discussing the current economic situation, I don't see it always as a bad sign.

Anyway, just my 2 cents..

[Brad Jones]

Arjay - If you take time to write some code that helps solve a poster's issue, then I suggest taking 15 more minutes to convert your post to an article with the code presented and submit it to the main site. If you tell Susan that it is a solution for someone in the forums, she can do a super quick review and post it "unedited" quickly to the article portion of the Codeguru site.

One of the issues I think we are trying to discuss in this thread is the concept of "correctness" and whether a thread should be corrected to maintain accuracy. Let's consider an example. Say a person posts the following:

I need to color the sky in my picture. What color do I use?

What is the answer?

Is it "blue"?

But.... that isn't really correct. The sky isn't blue. It appears blue many times, but really is that its true color, or just the color you happen to see - and only see at certain times? At night it -appears- black. In the evenings it can be all sorts of colors. Of course, when I said sky, did I mean on earth? On the moon, the "sky" is black.

I could go on.... and in many of our threads the conversation does go on, and on, and on....

If the original poster is a scientist looking at planet X in some other solar system, then the long discussion might be valid. But most of the time, this type of question is more likely to come from a 4 year old just trying to color a picture. For them, the answer is really pretty simple. 90% of the time it is "blue". Of course, you might ask a simple question of "is it day or night? Beyond that anything else said is noise.

It seems that many threads on this forum take a question like the color of the sky and turn them into discussions of varying quantites of oxygen versus other elements.....

In my opinion - I agree with Arjay in that we really have to take the original poster into consideration. We need to tone the answers to what they are looking to understand. Sometimes, we twist the truth a bit and say the sky is "blue" even though it really only *appears* blue. If we want to get into the finer accuracies of the question, best to spin off a new thread.

Brad!

(When I say blue, I mean Red 0, Green 128, Blue 255 with a hue of 140, saturation of 240, and lum of 120. --> I wouldn't want this thread to spin off into a discussion of what exactly did I mean when I said blue.... )

[Arjay]

IMO, don't rush the judgement. If you felt that it was useful to attach it. So be it. Don't look at the download count. You'll be amazed how many times I have got PMs, emails from people just saying "BTW, thanks for posting that information in that thread. It saved my day" and am sure there are many that didn't specifically email. Just feels good to know that somewhere someone finds some use for it.

I do get an occasional PM as well, just not too often. I'm more likely to continue if it seems folks are getting some benefit from them.

Hard to say if they are useless. Maybe not for the OP, but someone who bumps into the post later on could find the discussion useful. So, if iit's not insanely off-topic, like discussing the current economic situation, I don't see it always as a bad sign.

Just to be clear, I'm talking about the situation where a newbie poster has asked a relatively simple question, gets a reply and then another person replies which slightly contradicts the first reply. Then the thread goes off on a tangent debating the nuances of the 'solution'. The discussion may not be strictly off-topic but it may be irrelevant in terms of the original question (especially when the OP isn't participating). Perhaps it would be better to move the discussion to a new thread?

[Brad Jones]

The discussion may not be strictly off-topic but it may be irrelevant in terms of the original question (especially when the OP isn't participating). Perhaps it would be better to move the discussion to a new thread?

Absolutely!

[Arjay]

Cpu sent me a pm and asked me to respond in this thread.

I am actually most interested in thoughts regarding threads that ALREADY has one or more posts with "issues"....

I guess it depends on the degree of 'issues'. Using Brad's "sky is blue example", if the general answer of 'blue' gets the job done, then I'd say there's no point of going into more detail unless the OP asks for more detail (hue, saturation, etc.).

If a response is obviously wrong (like the sky is white), then it probably should be corrected in a non-personal manner.

Otherwise, more detail in the original thread may cause confusion so perhaps a separate thread would be in order (maybe titled "The benefits of 140 hue over 138")? A link to the new thread can be included in the original thread, so if any subsequent readers are interested they can follow the new discussion and others reviewing can get the quick answer.

[TheCPUWizard]

Since we hit 10 replies, and I have been mentioned....

#1 Thanks to everyone for sharing their thoughts.....

#2 While I think there is a general agreement on what should happen on a thread, I think the issue of "What to do AFTER a thread has 'encountered one of the issues' is still pretty open.

Yes, the moderators can split a thread, but their time (GREATLY APPRECIATED) is limited.

One of the aspects of a forum is the linear nature (based on time of post) in most views. Consider this in comparision to "News Groups" where it is always obvious which response a comment is in reference to. In that environment, one will often find "branches" that grow in different directions.

I am going to go through some recent threads later to find an actual example but for now consider:

Member #1: How can I accomplish "xxxx"

Member #2: Here is a sample using an std::vector. Code compiles, runs, and givens the correct answer, but is flawed because a temporary (which just didnt happen to get overwritten is use)

Members: {Comments on sample, but NOT catching the bug, rather covering differences in list , vs vector vs map....}

Member #1: Sorry this must be in "C"...

Member #3: Here is a sample again compiles and runs but has a memory leak

Members: That can be improved from an O(n^2) to an O(n) with the following code (new sample that works well and is valid and answers OP questions.

Now the posts from "member #2" and "member #3" are both still there, both are WRONG (ie contain hidden bugs)...but the OP has the answer from the final posts...

Do we simply allow the bug ridden code to remain forever without so much as a comment???

[kirants]

Do we simply allow the bug ridden code to remain forever without so much as a comment???

I would say yes to leave the buggy code, but sure leave a comment that the code is buggy. If we try to moderate it anymore, it becomes unmanageable. Do we trust Wikipedia for all the information in there? Not me. But it is a good starting point.

[TheCPUWizard]

I would say yes to leave the buggy code, but sure leave a comment that the code is buggy.

I am NOT suggesting editing the existing post...it is impossible for a non-moderator to do to anothers post anyway.

But the very act of adding a new reply, will likely trigger additional discussion about that particular aspect......

While I do not see this as a problem, it does account for a percentage of the "wandering" threads....

[Arjay]

I guess it depends on whether the buggy code is an obvious problem.

class Foo
{
  ...
};

int void Main( )
{
  Foe foo;
  foo.DoSomething( );

  return 0;
}

Sure, the code above won't compile, but it gets the point across and hopefully the user will catch the error (well all but the most beginning coder should catch the compiler error).

[TheCPUWizard]

I guess it depends on whether the buggy code is an obvious problem.<snip></snip>Sure, the code above won't compile, but it gets the point across and hopefully the user will catch the error (well all but the most beginning coder should catch the compiler error).

Typos, like that (unless pointed out immediately) are unlikely to cause problems even if totally ignored...

But...AS I posted...

Code compiles, runs, and givens the correct answer, but is flawed because a temporary (which just didnt happen to get overwritten is use)

It could be run for a long time without this error being exposed, but a simple re-org of the code would (possibly) break it.