>>If you still don't understand why my algorithm will work, do a google 
search
>>on "twos complement". Trust me...

I understood your algorithm and it will work fine if they have followed 
the 2nd approach, which is more likely because of its similarity to 
represent 8bit negative values.

>>1000011 does not equal -3, it equals -61. 
Yes, for an 8bit signed value. But i was just trying to say that the 
implementor might have also used a different representation(since its a 
7bit number) in which you take the Highest bit for sign & use only the 
rest of bits (6 in our case) for value.
thus 1000011 evaluates to negative 3.

But I am going ahead with your algorithm, since the representation is 
similar to 8bit negative values

Thanks & Regards
Praveen






"Ryan Ackley" <[EMAIL PROTECTED]>
31/07/2003 05:13
Please respond to "POI Developers List"

 
        To:     "POI Developers List" <[EMAIL PROTECTED]>
        cc: 
        Subject:        Re: HWPF queries

 

> The algorithm for signing a 7bit 2's complement value is cool, but i am
> confused as to  select which approach from the 2 given below.
>
> let me consider an example, say byte x = 3;  -> 00000011
> now x = -3 will be ->
> 11111101   (as per the usual representation of negative values)
>
> Ist approach:
> if we find the 7 bit value as                                    ->
> 0000011
> then check the 7th bit (Hi bit), its 0 so its a  positive 3
>
> if we find the 7 bit value as                                    ->
> 1000011
> then check the 7th bit (Hi bit), its 1 so its a  negative 3

1000011 does not equal -3, it equals -61. A negative twos complement is 
the
binary inverse of the positive number plus 1. The msb will always be 1 for 
a
negative number. By looking for the sign bit and simply filling the the 
8th
bit with a 1 is the same as filling it with a zero for a positive number.

If you still don't understand why my algorithm will work, do a google 
search
on "twos complement". Trust me...

Ryan


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
Mailing List:    http://jakarta.apache.org/site/mail2.html#poi
The Apache Jakarta POI Project: http://jakarta.apache.org/poi/



Reply via email to