[Logo] Function Point Analysis And Software Estimation Forum
  [Search] Search   [Recent Topics] Recent Topics   [Hottest Topics] Hottest Topics   [Members]  Member Listing   [Groups] Back to home page 
[Register] Register / 
[Login] Login 
Rounding ?  XML
Forum Index » Counting Related Questions
Author Message
Jan VOCHTEN


[Avatar]

Joined: 02/03/2012 16:08:45
Messages: 1
Location: Ekeren, Belgium
Offline

What if the Adjusted Function Point result is (for example) 100.1 ?
Do I round upwards to 101, or do I round down to 100 or do I keep the 100.1 ?
What do you do ?

Kind Regards

Jan VOCHTEN, CFPS (since 1999)
[Email]
RoyceEdwards



Joined: 08/02/2012 13:47:56
Messages: 4
Offline

Jan,

I would retain the full value. When it's translated to effort you don't want a round up error. The full value will look fine on a report. - Royce
caroldekkers


[Avatar]

Joined: 20/02/2012 05:16:20
Messages: 3
Location: USA
Offline

This is a great question and there is more to it than a question of pure accuracy.

When you report a number (like adjusted FP) that is based on components that are made up of integers (3,4,6, 4,5,7, 7,10,1, 5,7,10) it is statistical folly to even profess an accuracy (or precision!) that is closer than +/- 50% of the largest granularity number that goes into the mix. (I.e., +/- 50% of 15 is 7.5 so uFP is really no more accurate than +/- 7.5 FP).

When you then multiply the uFP by a two decimal point number, it is even more misguided to introduce a decimal point into the mix - this "pretends" we have some magical accuracy out of the multiplication process (ref: How to lie with statistics!) As if you can gauge even a single FP let alone a fraction of one.

It gets even worse when we are "estimating" thee function points from incomplete requirements which further introduces inaccuracies into the mix. We are doing our customers a disservice when we report anything but integer results.

No wonder when people say that their estimate is 100.1 FP that the project team rolls their eyes and management thinks we are wizards! We have to start acting our age when it comes to metrics and what we communicate!

This is a long answer to your short question, but I would never report the decimal place. Go with 100 FP and you'll find that the consequences lead to higher reliability on the part of the business and to your credibility.

Carol Dekkers, PMP, CFPS, CMC, CSM
Quality Plus Technologies, Inc.
phone 727 867 1949

Website: www.qualityplustech.com
Blog: http://musingsaboutsoftwaredevelopment.wordpress.com
[Email] [WWW]
faisalshakeel


[Avatar]

Joined: 05/02/2012 08:07:29
Messages: 7
Location: Pune, India
Offline

Hi Jan,

Let's use simple maths here. If the value is equal to or greater than .5, you have an advantage to round it up to 1 but if value is lesser than .5 then it's advisable to round it down. In your case you should take value as 100 not 101.

Hope this answers your question.

Regards,
Faisal Shakeel

Regards,
Faisal Shakeel
[Email] [Yahoo!]
pranithamudhiraj



Joined: 25/07/2017 11:31:07
Messages: 1
Offline

Test your understanding of Place value and rounding with these 10 questions. ... Going a bit deeper with our understanding of place value from 2nd grade. We'll apply this knowledge to learn to round things to the nearest 10 and 100.Rounding Numbers. What is "Rounding" ? Rounding means making a number simpler but keeping its value close to what it was. The result is less accurate, but ...Rounding a numerical value means replacing it by another value that is approximately equal but has a shorter, simpler, or more explicit representation; for ...
[WWW]
Steve Neuendorf



Joined: 02/02/2012 19:04:57
Messages: 2
Offline

I think there is a simple and logical answer. There is a real simple answer and it is to the question of why are you using adjusted function points when it fell from grace almost a decade ago (about 20 years too late IMNSHO)? That notwithstanding, as Carol points out you are constructing your UFP from a set of discrete numbers that are not even continuous. Assuming you have counted correctly so that accuracy is not an issue, the best possible precision for a typical count would be no less than + or - 5 FP. I would not be the least uncomfortable, and consider it an added service to my customers, (not to confuse them as the most recent post notes) to round everything up/down to the nearest 10. Don't fall into the "measure it with a micrometer; mark it with a grease pencil (soap); and cut it with an axe (torch) trap.

This message was edited 2 times. Last update was at 01/08/2017 22:06:08


Steve
 
Forum Index » Counting Related Questions
Go to:   
Powered by Function Point Modeler © FPM Team