[vwnc] Discussion on ScaledDecimal (FixedPoint in VW) ...

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

I agree that it does something useful, I just don't think that what it does is fixed point.

At 06:02 PM 2009-12-16, Dennis Smith wrote:

I disagree -- if you actually USE FixedPoint (I don't care what you call it) -- it does exactly what is
required for accounting apps -- and I have been working with Accounting and Payroll for about 12 years now.

We do large A/R, A/P, G/L and Payroll apps -- and we need exactly what FixedPoint provides.

At various points one has to Round or Truncate (or Ceil) based on what you are doing, and you need
control over when it occurs.  Having it arbitrarily round at each operation would make it useless to us and
we would then have to create out own.

Whoever designed this in VW new what they were doing.


Alan Knight wrote:

It seems to me that the entire discussion here boils down to the VisualWorks FixedPoint class being poorly named. What it does is useful functionality, but it's not really fixed point, and there isn't anything in VisualWorks that does do real fixed point as far as I know. As John Brant said, one can just use fractions for indefinite precision, and that's really all that class is - a wrapper around fractions. To do real fixed point,  you just use integers, so a wrapper around integers would accompish the same sort of thing. And then the only worry would be all the people using the existing class who weren't expecting it to go from essentially infinite precision to extremely limited precision. Or we could name it ReallyFixedPoint :-)


--
Alan Knight [|], Engineering Manager, Cincom Smalltalk
[hidden email]
[hidden email]
http://www.cincom.com/smalltalk

_______________________________________________
vwnc mailing list
[hidden email]

http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
 

-- 
Dennis
Smith                
   
        
         +1 416.798.7948
Cherniak Software Development Corporation   Fax: +1
416.798.0948
509-2001 Sheppard Avenue East       
[hidden email]
Toronto, ON M2J
4Z8             
<a href="sip:dennis@CherniakSoftware.com">
sip:dennis@...
Canada  
        
        
        

http://www.CherniakSoftware.com
Entrance off Yorkland Blvd south of Sheppard Ave east of the
DVP

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

I disagree.
The result of a computation is some value with a fixed decimal point.
If that value is 1/3 and you then display it in normal form you get
    0.333333333333333333333333333- forever
but then it says its a Fixed Decimal Point at 4 say, so you really get
    0.3333
If you then say, no I want scale 5 you get
    0.33333
who said you should only get
    0.33330
just because you changed the scale??

Alan Knight wrote:

It seems to me that the entire discussion here boils down to the VisualWorks FixedPoint class being poorly named. What it does is useful functionality, but it's not really fixed point, and there isn't anything in VisualWorks that does do real fixed point as far as I know. As John Brant said, one can just use fractions for indefinite precision, and that's really all that class is - a wrapper around fractions. To do real fixed point,  you just use integers, so a wrapper around integers would accompish the same sort of thing. And then the only worry would be all the people using the existing class who weren't expecting it to go from essentially infinite precision to extremely limited precision. Or we could name it ReallyFixedPoint :-)

--

Alan Knight [|], Engineering Manager, Cincom Smalltalk

[hidden email]

[hidden email]

http://www.cincom.com/smalltalk

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
  

-- 
Dennis Smith                 		         +1 416.798.7948
Cherniak Software Development Corporation   Fax: +1 416.798.0948
509-2001 Sheppard Avenue East        [hidden email]
Toronto, ON M2J 4Z8              <a class="moz-txt-link-freetext" href="sip:dennis@CherniakSoftware.com">sip:dennis@...
Canada			         http://www.CherniakSoftware.com
Entrance off Yorkland Blvd south of Sheppard Ave east of the DVP

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Am 16.12.2009 um 23:58 schrieb Dennis Smith:

> Hope you don't get an audit by a knowledgeable accountant -- they  
> won't accept it. Floats (and Doubles) have issues with accuracy in  
> accounting -- that is why we have various forms of FixedPoint  
> (ScaledDecimal).  Most databases provide a storage and even a  
> computational form of  Decimal values to handle those cases.

Yep. My mistake was to think FixedPoint was based on a fixed point  
representation of the value. Undoubtedly an implementation based on  
fractions is superior. I just didn't realize how the class was  
implemented and took its name for real.

I admit I'm not specialized in accounting software. For my purposes  
Doubles were sufficient. I would probably use FixedPoint now that I  
finally got around to read the entire class comment ;-)

Sorry for the confusion.

Andre

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Joachim Tuchel wrote:
> Wouldn't ScaledDecimal be a good name?
> Maybe not, since a ScaledDecimal is the equivalent of FixedPoint in
> other Smalltalk dialects...

ScaledDecimal might indeed be a good name. In VA, ScaledDecimal is an
actual fixed point class, and the current proposal is that GemStone's
"real" fixed point class will be named ScaledDecimal. For more
information on the current thinking at GemStone, see James Foster's blog:
http://programminggems.wordpress.com/2009/12/16/scaleddecimal/


Regards,

-Martin
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

On Dec 16, 2009, at 2:40 PM, Alan Knight wrote:

Well put. Smalltalk FixedPoint (ScaledDecimal in other dialects) are  
really just FractionWithPrintPrecision numbers.

We'll never get rid of it, or "correct" it (for some definition of the  
word correct). It is after all an integrated literal type (even though  
Fractions aren't) and documented in the ANSI spec I believe.

Someone knew what they were doing as far as implementing a type of  
numeric object that retains a display precision, and didn't know what  
they were doing when they were choosing names.

--
Travis Griggs
Objologist
"The project was so plagued by politics and ego that when the  
engineers requested technical oversight, our manager hired a  
psychologist instead." -- Ron Avitzur

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Travis Griggs wrote:

On Dec 16, 2009, at 2:40 PM, Alan Knight wrote:

  

It seems to me that the entire discussion here boils down to the  
VisualWorks FixedPoint class being poorly named. What it does is  
useful functionality, but it's not really fixed point, and there  
isn't anything in VisualWorks that does do real fixed point as far  
as I know. As John Brant said, one can just use fractions for  
indefinite precision, and that's really all that class is - a  
wrapper around fractions. To do real fixed point,  you just use  
integers, so a wrapper around integers would accompish the same sort  
of thing. And then the only worry would be all the people using the  
existing class who weren't expecting it to go from essentially  
infinite precision to extremely limited precision. Or we could name  
it ReallyFixedPoint :-)
    

Well put. Smalltalk FixedPoint (ScaledDecimal in other dialects) are  
really just FractionWithPrintPrecision numbers.

We'll never get rid of it, or "correct" it (for some definition of the  
word correct). It is after all an integrated literal type (even though  
Fractions aren't) and documented in the ANSI spec I believe.

Someone knew what they were doing as far as implementing a type of  
numeric object that retains a display precision, and didn't know what  
they were doing when they were choosing names.
  

If you take that point of view, is Integer misnamed because  1 / 3 doesn't produce an integer?
In fact
    1 / 3 * 2
also does not produce an integer.
Why should FixedPoint be any different?  Why should
    1s / 3 * 2
produce 0s as a result??

--
Travis Griggs
Objologist
"The project was so plagued by politics and ego that when the  
engineers requested technical oversight, our manager hired a  
psychologist instead." -- Ron Avitzur

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
  

-- 
Dennis Smith                 		         +1 416.798.7948
Cherniak Software Development Corporation   Fax: +1 416.798.0948
509-2001 Sheppard Avenue East        [hidden email]
Toronto, ON M2J 4Z8              <a class="moz-txt-link-freetext" href="sip:dennis@CherniakSoftware.com">sip:dennis@...
Canada			         http://www.CherniakSoftware.com
Entrance off Yorkland Blvd south of Sheppard Ave east of the DVP

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Dennis Smith wrote:
>
> Travis Griggs wrote:

>> We'll never get rid of it, or "correct" it (for some definition of the  
>> word correct). It is after all an integrated literal type (even though  
>> Fractions aren't) and documented in the ANSI spec I believe.

I think VA Smalltalk implements ScaledDecimal as a true fixed point
number (http://en.wikipedia.org/wiki/Fixed-point_arithmetic). Also, the
ANSI spec says that the <scaledDecimal> protocol "Provides a numeric
representation of fixed point decimal numbers." The VW FixedPoint class
does not implement a fixed point number.

>> Someone knew what they were doing as far as implementing a type of  
>> numeric object that retains a display precision, and didn't know what  
>> they were doing when they were choosing names.
>>  
> If you take that point of view, is Integer misnamed because  1 / 3
> doesn't produce an integer?
> In fact
>     1 / 3 * 2
> also does not produce an integer.

As defined by the ANSI standard: 1/3 produces a fraction and the
fraction (1/3) * 2 produces another fraction.


> Why should FixedPoint be any different?  Why should
>     1s / 3 * 2
> produce 0s as a result??

Because that's what a fixed point calculation (with no fractional
digits) would produce. The ANSI standard says that 1s / 3 is evaluated
as "1s / 3s" and that / is a flooring division operation so 1s / 3 is 0s.

Dennis, I'm not saying that what VW provides isn't useful. I'm saying
that it isn't a fixed point number and doesn't provide the
<scaledDecimal> protocol from the ANSI standard.


John Brant
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

John Brant wrote:

Dennis Smith wrote:
  

Travis Griggs wrote:
    

We'll never get rid of it, or "correct" it (for some definition of the  
word correct). It is after all an integrated literal type (even though  
Fractions aren't) and documented in the ANSI spec I believe.
      

I think VA Smalltalk implements ScaledDecimal as a true fixed point
number (http://en.wikipedia.org/wiki/Fixed-point_arithmetic). Also, the
ANSI spec says that the <scaledDecimal> protocol "Provides a numeric
representation of fixed point decimal numbers." The VW FixedPoint class
does not implement a fixed point number.

  

Someone knew what they were doing as far as implementing a type of  
numeric object that retains a display precision, and didn't know what  
they were doing when they were choosing names.
  
      

If you take that point of view, is Integer misnamed because  1 / 3
doesn't produce an integer?
In fact
    1 / 3 * 2
also does not produce an integer.
    

As defined by the ANSI standard: 1/3 produces a fraction and the
fraction (1/3) * 2 produces another fraction.


  

Why should FixedPoint be any different?  Why should
    1s / 3 * 2
produce 0s as a result??
    

Because that's what a fixed point calculation (with no fractional
digits) would produce. The ANSI standard says that 1s / 3 is evaluated
as "1s / 3s" and that / is a flooring division operation so 1s / 3 is 0s.

Dennis, I'm not saying that what VW provides isn't useful. I'm saying
that it isn't a fixed point number and doesn't provide the
<scaledDecimal> protocol from the ANSI standard.
  

What the ANSI standard says and what is of value for accounting appear to be two different things.
At the very least we must have truncation not rounding -- nothing else rounds.

Look at Java Decimal type
    Decimal abc (2)
is a 2 scale value which retains 5 points of precision -- I wonder why?  You can override that, but that
is the default -- well thought through.

As s/w engineers we should not be defining how accounting needs to work.

John Brant
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
  

-- 
Dennis Smith                 		         +1 416.798.7948
Cherniak Software Development Corporation   Fax: +1 416.798.0948
509-2001 Sheppard Avenue East        [hidden email]
Toronto, ON M2J 4Z8              <a class="moz-txt-link-freetext"
 href="sip:dennis@CherniakSoftware.com">sip:dennis@...
Canada			         http://www.CherniakSoftware.com
Entrance off Yorkland Blvd south of Sheppard Ave east of the DVP

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc