SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    Re: iSCSI: Decimal encoding - why 64 bits ?



    If you fear about extended precision arithmetic you should carefully look at
    what key=values are.
    Precluding the use of large numerical values by forbidding their encoding is
    not a good way of considering this - we have also base64 for large numbers
    (and hex) and if you have to have a use for larger numbers you have to have
    the arithmetic to handle it.
    
    As for the second subject - forbidding the use decimals for binary strings -
    we never discussed it or agreed on this.
    
    I wonder how natural it will feel for somebody to have and IP address
    encoded exclusively in hex, or a TargetPortalGroup appearing as decimal in a
    directory and having to be transliterated.
    
    Julo
    
    ----- Original Message -----
    From: "Bill Studenmund" <wrstuden@wasabisystems.com>
    To: "Julian Satran" <Julian_Satran@il.ibm.com>
    Cc: <ips@ece.cmu.edu>
    Sent: Wednesday, July 03, 2002 12:22 AM
    Subject: RE: iSCSI: Decimal encoding - why 64 bits ?
    
    
    > On Tue, 2 Jul 2002, Julian Satran wrote:
    >
    > > It was never supposed to be removed. Many values are passed around as
    > > decimal.
    > > We can't make any progress if we keep hitting the same things
    > > again-and-again after a decent consensus has been reached.
    >
    > The question is does the draft reflect the concensus that all the
    > discussion participants thought they achieved?
    >
    > At least one other person had the same impression I did about the past
    > discussion. i.e. we thought we HAD achieved concensus, and yet the draft
    > does not reflect that discussion.
    >
    > > And none of you has brought an argument that was not heard and dismissed
    > > before.
    >
    > When were these arguements dismissed? While I recall a lot of disagreement
    > on points, I don't recall a sound dismissal on technical grounds.
    >
    > > Remember we moved from unlimited length decimal to 64 bit to alleviate
    > > implementer fears.
    >
    > Julian, those fears were for something else. Those fears were for how do
    > you deal with extended-precision math when reading a large number.
    >
    > These concerns are that decimal encoding of binary strings suffers from
    > many of the problems that base64 had for numbers - the need to perform
    > arithmetic to byte-string conversion (i.e. hotns() and htonl() & friends).
    >
    > Now that the text explicitly states that the string size is in whole
    > bytes, things aren't as weird as before. But it's still messy. I'll post a
    > seperate message on this.
    >
    > Take care,
    >
    > Bill
    >
    
    


Home

Last updated: Mon Jul 08 19:18:53 2002
11186 messages in chronological order