CDRLabs.com

KCK wrote:Namely, some time ago on a networking forum I reported transfer speeds in MB/s, adding that 1 MB = 10^6 B in that context. This started a heated discussion with a crowd of computer scientist, who claimed that such usage was not supported by any standard.

That is surprising, to me. I've always associated the SI standard 1 M = 10^6 usage with networking, so I wouldn't have expected the confusion there. But, as you say, it does depend on context, and that is what makes it all so confusing. Perhaps we should do our part to fight the confusion by using the kibi, mebi, etc. prefixes when appropriate. Of course, since they aren't widely known among many users, that would just create its own confusion...

cfitz

I think this is a good idea. I've had an uneasy feeling about the tacit abuse of decimal multipliers for binary factors since I first encountered them. Completely apart from marketing hype, this is a problem that will increase as storage capacity goes up, compounded by the problem of hybrid units mentioned in the IEEE draft:
- k(i)B: 2% This is nothing to worry about
- M(i)B: 5% Some care required
- G(i)B: 7% Here creative marketing comes into its own.
- T(i)B: 10% worse
Of course I'm a physicist, but most physicists have more than casual contact with computers, so I don't think this disqualifies me. The IEEE draft is new to me, but I like the idea. The units are even melifluous- no hard G's or R's. I don't know how long its been kicked around- its dated April, 2002, so there must be comments, and I hope it gets adopted.

I'll start using MiBs etc. as of now.

G

The revolution starts here, huh? See, I told you physicists were offended by the computer scientists' bastardization of the SI system.

Okay, I'll resolve to do the same, and try to be part of the solution rather than part of the problem. If people have questions about it, I will refer them to this thread.

cfitz

P.S. I personally prefer the hard consonants. "Jigahertz?" Please! I shudder just thinking about it.

cfitz wrote:The revolution starts here, huh? See, I told you physicists were offended by the computer scientists' bastardization of the SI system.

Right! Bring out the cannons (small i's) and the little red/green books (or whatever colour the IEEE is using).

cfitz wrote:"Jigahertz?" Please! I shudder just thinking about it.

Same here- and GiB pronounced as JeebieBytes ?.

G

For future references, here is more info on prefixes for binary multiples.

These prefixes with names/symbols kibi/Ki, mebi/Mi, gibi/Gi, tebi/Ti, pebi/Pi and exbi/Ei were approved as an IEC International Standard in December 1998 by the International Electrotechnical Commission (IEC). The current complete citation is IEC 60027-2, Second edition, 2000-11, Letter symbols to be used in electrical technology - Part 2: Telecommunications and electronics, as given in

http://physics.nist.gov/cuu/Units/binary.html

On 10 December 2002, the IEEE-SA Standards Board Review Committee approved P1541/D5 (SCC14) Trial-Use Standard for Prefixes for Binary Multiples. Its contents apparently coincide with the draft discovered by cfitz:

http://www.ieee802.org/secmail/pdf00106.pdf

Here note that trial-use standards normally become full-use standards after two years, unless they meet sufficiently strong opposition.