RE: iSCSI: Decimal encoding - why 64 bits ?

[Paul Koning <ni1d@arrl.net>]

>>>>> "Bill" == Bill Studenmund <wrstuden@wasabisystems.com> writes:

 Bill> On Tue, 2 Jul 2002, THALER,PAT (A-Roseville,ex1) wrote:
 >> Bill,
 >> 
 >> I also thought that the result of the reflector discussion was
 >> agreement that decimal would only be used for numbers and not be
 >> used for strings that are sometimes longer than the 64 bit limit
 >> like CHAP and SRP strings. That seems to be partially implemented
 >> in that the Kerbos and SPKM sections call for large-binary-values,
 >> but it was done incompletely.

 Bill> I looked at the list archive, and
 Bill> http://www.pdl.cmu.edu/mailinglists/ips/mail/msg10319.html is a
 Bill> comment by David indicating he saw concensus for not using
 Bill> base64 for numbers.

 Bill> I didn't find the thread part where I thought we agreed that we
 Bill> would use decimal only for numbers (and not for small binary
 Bill> strings).

That certainly is the conclusion I find most reasonable.  The allowed
encoding should be a property of the thing encoded.  We have a pile of
integers, and we have a few bitstrings.  The integers need to allow
decimal encoding; the bitstrings should NOT allow it.

	paul