Bankers Rounding
-
By the way, this could have been a simple debate on the merits of banker's rounding (there's plenty of room for debate in that area, some of which I'd probably side with you on), where it is most applicable, and how to accomplish it, but if you are going to be an ass about something, it helps to have the facts on your side. Hell, even if you were right (you're not), there's still no reason for comments like the following in technical forum, especially when debating people such as Lynn.
Author of this conception missed something important on 5th or 6th grade in school.
Someone who wrote this definitely missed couple of important things in school.
I stated that whole idea of Bankers Rounding is stupid. And I can prove it.
"Too much proud" problem?
That's where brilliant software built on something like Bankers Rounding fails and the boring academic guy comes to fix it and save the company from Revenue Department charges (before him several people had to go every day through every bloody Invoice Report and manually correct odd numbers).
Earlier in the thread, you stated "Lynn, I will admit anything you can prove. Can you?" Maybe it's time that you back that up with action.
Everyone makes mistakes. I probably make them more often than most people. Owning up to them is a "best practice" in my book, as one, it's honest, and two, once admitted, they typically are forgotten quite quickly.
As an aside, it should be noted that .NET actually implemented round-half-even (banker's rounding) in their Math.Round() function initially (so there are a ton of people using the method, whether they know it or not), but recently allowed it to be overwritten, not because it was bad, but because the behavior was unexpected. I agree with this decision, as there are times when you don't want to use it.
-
David, you put too many words but did not answered my questios.
One of them:
Which day the midnight event belongs? Previous or next?
Forget about bankers. Answer for yourself.
_____________
Code for TallyGenerator -
Okay, I can't stay out of it any longer. David, you have put up a good argument, but it is time to end it. No matter how much we try to explain this to Sergiy, he will continously find a way to twist it. He obviously has to be right even when he isn't.
Sergiy, by current definition, midnight is the next day, but I have to ask, which midnight are you talking about. You never stated which timezone you were looking at the clock. Midnight in New Zealand isn't midnight in the US or UK.
-
You promissed - stay out.
When you have learnt not to take personal point in discussions you'll be allowed to join.
We'll definitely find the common groung with David without your "help".
_____________
Code for TallyGenerator -
You changed the subject when you brought up midnight. We are no longer talking about the bankers round function.
-
-
Huh? This started as a topic on a Bankers Rounding function, and then you ask what day do you consider a transaction occuring when it occurs at midnight? What does one have to do with the other? I'll tell you what, absolutly nothing, they are completely unrelated.
Point in fact, previous employer is a world-wide company with offices in the United States, UK, France, Germany, Hong Kong, Australia (to name a few). Business rule used: Any transaction received on or after 6:00 PM Mountain Time (Day Light or Standard depending on the season) was treated as being received the next business day. This meant that some transactions received during the local day in Australia and Hong Kong were split across days at the corporate office.
Now if you can tell me what that has to do with validity or use of the Bankers Rounding function, you have a twisted sort of logic that I simply can not fathom.
-
Ever wish you had not posted something? (Of course you have).
People, enough already.
Lynn, in his last post said
"This started as a topic on a Bankers Rounding function"
Actually, (and I'm sorry to disagree with him as his initial post was bang on the money), it was more along the lines of "Why does this algorithm not work?"
David
"this could have been a simple debate on the merits of banker's rounding"
Ditto, really.
Sergiy
You solved the problem, but continue to argue about the validity of the method. Have you never had to do something, which in your own opinion, was against the best interest of your end user? I was asked to come up with a bankers rounding function in T-SQL, not discuss the merits/shortcomings of the specific rounding method. It is a defined standard,
(from wikipedia)
"The round-to-even method is the same except that numbers exactly half-way between the possible outcomes are sometimes rounded up—sometimes down."
You also said
"That's why I strongly opposed the statement in initial post mentioned "large sets of scientific or statistical data". Using Bankers Rounding for these purposes is absolutely inappropriate."
If you look at the initial post, it's a quote from wikipedia, as credited. This gives it the same provenance as any thing posted on any website. Including this one. So therefore I can take anything you post as valid as anything else by anyone else. Anywhere
The problem ("Why does this algorithm not work") is solved, let it go.
Dave J
-
Dave J.
Okay, I agree.
Lynn
-
OK, Lynn, as soon as you admit you were wrong and you can prove something to me (no point to ask for apologi, I guess), lets continue.
You said, midnight event belongs to the next day.
I agree.
Next question:
What you'd define as "midnight event"?
When midnight event must take place to be defined as such?
_____________
Code for TallyGenerator -
I WANT MORE!
This thread is hilarious.
Can't we discuss negative numbers and rounding now?
N 56°04'39.16"
E 12°55'05.25" -
I have nothing to admit to being wrong about, nor do I have reason to apologize.
As you have stated in several posts in this thread, you haven't answered my question. Still waiting to see what the Bankers Round function has to do with when one recognizes a transaction that occurs at midnight has to do with each other.
-
If you entered the argument again and you did not admit you're wrong what's the point then?
Move to personal insults when you fail again?
Still waiting to see what the Bankers Round function has to do with when one recognizes a transaction that occurs at midnight has to do with each other
Easily.
According to that logic midday (0.5 days) belongs to "pm" time. Right?
Right.
So, "am" hours:
0:00, 1:00; 2:00; 3:00; 4:00; 5:00; 6:00; 7:00; 8:00; 9:00; 11:00; 11:00;
and "pm" hours:
12:00, 13:00; 14:00; 15:00; 16:00; 17:00; 18:00; 19:00; 20:00; 21:00; 22:00; 23:00.
Same story with mass, length, height, etc.
Value whih reflects a division between 2 intervals (so colled "precise" number) belongs to the next interval.
There are no "between intervals" numbers. It's nonsense in real worlds.
This is something from school book, not sure which grade in American schools.
Because money is an abstraction to reflect real world (it another measurement of things, it shows not mass or length but cost), same rule must be applied here as well.
$0.5 belongs to the second half of a dollar as well as 0.5kg belongs to second part of a kilogram and 12:00 belongs to the second half of the day.
Not to any "between halves" numbers.
So, it must be rounded to the same value as the rest of interval.
Always.
End of story.
The part I did not mention here is about "precise" numbers.
If public will express some kind of interest I could tell another story from school book, this time about so called "precise numbers".
For the beginning you could answer the question:
What does it means when they say: "Average cost of a household in New Zealand is 245 thousand USD" (value is false, just made it up).
Give a full explanation of that expression, please.
_____________
Code for TallyGenerator -
I have never resorted to insults (unless you you are meaning my comment about a type a personality, taht was not an insult but an observation on my part), although the same can't be said of your posts.
I am not going to get into an academic argument with you. Simply put, there is a valid reason for implementing a Bankers Round (or round to even) function, and it does of an application in accounting and tax applications. Even then, you may still run into a penny (0.01) off error, but every one of those errors simply means implementing a BUSINESS RULE as to where the left over penny gets assigned.
If you wish to continue the academic argument, please find someone else who is interested in such an argument. The real world of business is made up of comprimises, not the absolutes of academia.
As Dave J. said earlier, time to let it go.
-
You are a brainless moron.
Please don't take it as an insult, it's just an observation on my part.
Happy with that?
(It does not mean it's my opinion, it's just a reflection of your words)
Simply put, there is a valid reason for implementing a Bankers Round
No, there is not.
Nobody can show me one.
If you wish to continue the academic argument, please find someone else who is interested in such an argument.
You mean who is smart enough and can understand this stuff?
_____________
Code for TallyGenerator
Viewing 15 posts - 61 through 75 (of 373 total)
You must be logged in to reply to this topic. Login to reply