JMP Mail Archive: JMP> Re: rounding up in jobKOctetsRequested(48) vs.

JMP> Re: rounding up in jobKOctetsRequested(48) vs.

Tom Hastings (hastings@cp10.es.xerox.com)
Tue, 6 May 1997 11:19:26 PDT

At 17:27 05/05/97 PDT, Harry Lewis wrote:
>Tom, I'm trying to understand why you chose two different ways to represent
>kOctets.
>See jobKOctetsRequested(48) v.81 pg. 42 where 1-1024 is represented as 1 vs.
>jobKOctetsTransferred(49) and jobKOctetsCompleted(50) v.81 pg. 43 where 1-1023
>is represented as 1 vs.
>
>Is this just a typo?
>
>Harry Lewis - IBM Printing Systems
>
>

Oops! The latter two are typos.

All three should be the same as follows:

Actual octets Rounded K value
------------- ---------------
0 0
1-1024 1
1025-2048 2
2049-3072 3
3073-4096 4
etc.

In other words, if the actual value is an integral number of K (1024) then
the number of K is the value. If the actual value is more than an integral
number of K, then round up to the next higher number of K.

This can be computed simply as: 1 + IntegerPartOf((ActualOctets-1)/1024)

Is it sufficient if I change the descriptions of
jobKOctetsTransferred(49) and jobKOctetsCompleted(50) on page 43 to:

The agent shall round the actual number of octets [transferred/completed]
up to the next higher K. Thus 0 octets shall be represented as 0,
1-1024 octets shall be reprsented as 1, 1025-2048 octets shall be represented
as 2, 2049-3072 octets shall be represented as 3, etc.

Or do we need a more rigorous mathematical expression, such as the one above
or is there a better way to say this in words?

Thanks,
Tom