Clarify between decimal bytes and binary bytes

Jack Cole jack.cole at IEEE.org
Thu Jul 8 06:55:22 PDT 2010


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1007080_f.htm">HTML-formatted message</a>

 FWIW, the introduction to IEEE 1541 and IEEE 1541 itself may be useful to
your pursuit. 1541-2002
IEEE Standard for Prefixes for Binary Multiples
http://ieeexplore.ieee.org/servlet/opac?punumber=5254929
Modern computers use binary logic for computation and addressing, and binary
logic inevitably leads to
addresses expressed in binary arithmetic. The size of such an address space
is inevitably a power of two.
Thus, when computer memories and disks were tiny (in terms of capacity), it
became common practice to
use “kilo” as a prefix denoting multiplication by 1024 (= 210). In the 1960s
and 1970s this created no
problem because there is not much difference between 1000 and 1024, and
within the community of persons
who used computers, everybody understood what was meant. Thus decimal
prefixes came to be applied on
the grounds that it would have been senseless, in the computer field, to
interpret them as anything other than
binary multiples, i.e., it would have seemed illogical to size a small
memory in multiples of 1000 when the
size of the address space was 1024. As the capacity of memories and disks
has grown larger, the issue of
correspondence with the size of the address space became less important than
the issue of total capacity
requirements. In addition, the disparity between binary and decimal
multiples is larger with the larger
prefixes. Data storage specialists now work with terabytes. If one purchases
a terabyte of storage, can one
store 1012 bytes or 240 bytes? The difference is roughly 10%.
Personal computers have become ubiquitous in the twenty-first century, and
the use of decimal prefixes
where binary multiplication is intended causes real confusion. Most computer
users today are not specialists.
They know that a kilometer is 1000 meters and, having no familiarity with
powers of 2, assume that a
kilobyte is 1000 bytes. The result is confusion and occasional
misunderstanding. This IEEE standard defines
new prefixes for binary multiples and thereby makes precise and unambiguous
communication possible. A
similar standard, IEC 60027-2 [B1], has already been adopted by the
International Electrotechnical
Commission.
On Wed, Jul 7, 2010 at 12:37 PM, Kevin D Butt <kdbutt at us.ibm.com> wrote:
>
> I believe that the time has come to clear up confusion in the industry
> about the units used in T10 standards.  There should be a clause added in
> each of the standards describing the difference between decimal bytes
(e.g.,
> MB = 10^6) and binary bytes (e.g., MiB = 2^20).  There has been this
> definition in IEC for since Dec 1998.
>
> References:
> http://physics.nist.gov/cuu/Units/binary.html
> http://en.wikipedia.org/wiki/Binary_prefix
>
> I hesitate to bring in a solo proposal to modify SSC-4 or even SPC-4 when
> this issue should apply to all standards.
> Does anybody disagree that T10 should clarify the units being used in our
> standards?
> What is the preferred manner in which we should proceed to make sure this
> clarification is added to each standard?
>
> Thanks,
>
> Kevin D. Butt
> SCSI & Fibre Channel Architect, Tape Firmware
> MS 6TYA, 9000 S. Rita Rd., Tucson, AZ 85744
> Tel: 520-799-5280
> Fax: 520-799-2723 (T/L:321)
> Email address: kdbutt at us.ibm.com
> http://www-03.ibm.com/servers/storage/



More information about the T10 mailing list