Re: iSCSI - decimal coded binary strings - a proposed resolution

["Julian Satran \(Hotmail\)" <Julian_Satran@il.ibm.com>]

Paul,

There seems to be something missing.
There is no decimal coding for numbers over 2**64.
And there no implied length for strings.
Where is the complexity?

Julo
----- Original Message ----- 
From: "Paul Koning" <ni1d@arrl.net>
To: <Julian_Satran@il.ibm.com>
Cc: <ips@ece.cmu.edu>
Sent: Tuesday, July 09, 2002 5:08 PM
Subject: Re: iSCSI - decimal coded binary strings - a proposed resolution


> >>>>> "Julian" == Julian Satran <Julian_Satran@il.ibm.com> writes:
> 
>  Julian> Paul - I think we are talking again about two different
>  Julian> thinks:
> 
>  Julian> numerical values that could be more than 2**64 - I would not
>  Julian> forbid it there
> 
>  Julian> bit strings that could be longer than 64 bits - I find it
>  Julian> acceptable
> 
>  Julian> The new text could be:
> 
>  Julian> decimal-constant: an unsigned decimal number - the digit 0 or
>  Julian> a string of 1 or more digits starting with a non-zero digit.
>  Julian> Decimal-constants are used to encode numerical values or
>  Julian> binary strings. Decimal constants can be used to encode
>  Julian> binary strings only if the stringlength is explicitly
>  Julian> speci-fied. There is no implicit length for decimal
>  Julian> strings. This encoding MUST NOT used for numerical values
>  Julian> equal or greater than 2**64 or binary strings that could be
>  Julian> longer than 64 bits.
> 
> There currently are no numeric values >= 2**64.  But in any case, I
> don't like the idea there either.  What I proposed makes the encoding
> an attribute of the parameter -- either it can always be encoded in
> decimal, or it never can be.
> 
> What you're proposing does that for binary strings.  But for numeric
> values, the encoding permitted would depend on the particular value.
> I don't see any sense in that extra complexity.
> 
>     paul
>