Husky71

Kelvin bits per second?

8 posts in this topic

My upload speed test result was reported with units of Kbps.

 

I know the intent was to state the speed in kilo bits per second, but per SI unit rules a capital K is the unit for temperature, kelvin.  The ten to the third power multiplier prefix kilo is abbreviated using a lower case k.  See http://physics.nist.gov/cuu/Units/

 

And to be strictly correct, the internationally standardized abbreviation for binary digit is bit.  International standards bodies do not recognize b as a further abbreviation for bit.

Interesting factoid, John Tukey, who worked at Bell Labs, first suggested the word bit as an abbreviation for binary digit.  Claude Shannon first used the word in publication in his paper A Mathematical Theory of Communications.

 

And to go one step further, SI unit usage and grammar rules frown on the use of p to represent per in a unit formulation.

 

So strictly the unit would properly be kbit/s or kbit * s^-1

CA3LE likes this

Share this post


Link to post
Share on other sites

You're obviously 100% correct.  I took a few minutes to make global corrections to the site.  If you see references on the main site (outside of the forums) to Kbps or Mbps please bring it to my attention.  I should have caught all the occurrences but I'm human... hence the reason it needs to be corrected in the first place.

 

I appreciate you sharing your knowledge and pointing this out.  TestMy.net has always been built in your feedback.  I'm just the person typing, you're the ones building this with your feedback.  Can't tell you how many times my users have corrected me.  I'll stand corrected each time and smile because it's made my site better.  It's what has made TMN what it is today.  

 

I'm often wrong but I try to be quick at making corrections and admitting it.  

 

Hundreds of millions of visitors, maybe upwards of 1/2 a billion and nobody has said a word about this detail.  

wenfinger likes this

Share this post


Link to post
Share on other sites

This is an interesting point . . .

 

Yet I wonder if it falls under the category of "accepted vernacular", even in the IT field.  

 

Similar to a medical writer arguing with a bench scientist over the use of criteria/data vs criterion/datum in publication.  While technically accurate to use the Greek-based criterion/datum for a singular point, it has become common-place and, dare I say, accepted to say data (or even agenda. Do we stop to think about that?)

 

Just my $0.02

 

HT to CA3LE for professionalism and willingness to be fluid and adaptable with making TMN better. :D

Share this post


Link to post
Share on other sites

 Considering back 15 or even 20 years ago as a whole, we all schooled one another on proper distinctions between kilobit and kilobytes ect.

At this point I'm inclined to take @wenfingers thought, until we can run this back to it's true roots, wherever that may reside.

 

I'd like to get @nanobots input on this one.

Share this post


Link to post
Share on other sites

Odd that for frequency, it's Hz, kHz, MHz and GHz. Only the "k" is lower case. The "H" in Hz is upper case because it's derived from a man's name: Hertz.

 

the rules in the standards are suffocatingly detailed

Share this post


Link to post
Share on other sites

As someone who works in the industry developing recommendations and standards, I'm happy to provide references to the various internationally recognized references where the various terms are defined or codified.

 

There are all sorts of references out there that people point to from time to time that are not authoritative sources on the topic, meaning that they are not traceable to a standards body such as BIPM, IEC/ISO, ITU, etc.  I've seen mistakes in those documents, which unfortunately only continue to further propagate misuse and confusion.

 

Share this post


Link to post
Share on other sites
On 5/1/2017 at 7:00 AM, mudmanc4 said:

 Considering back 15 or even 20 years ago as a whole, we all schooled one another on proper distinctions between kilobit and kilobytes ect.

At this point I'm inclined to take @wenfingers thought, until we can run this back to it's true roots, wherever that may reside.

 

I'd like to get @nanobots input on this one.

 

Using Kbps / kbps is misleading, as we're *actually* referring to KiBps. (One KiB = one kibibyte, which is 1024 bytes, one KB = one kilobyte, which is 1000 bytes.)

 

The definitions of the prefixes are pretty lame, but the SI version is the following:

 

Multiplier   Abbr   Prefix (Example)
1                   NONE  (bytes)
1000         k      kilo  (kilobytes)
1000^2       M      mega  (megabytes)
1000^3       G      giga  (gigabytes)
1000^4       T      tera  (terabytes)
1000^5       P      peta  (petabytes)
1000^6       E      exa   (exabytes)
1000^7       Z      zetta (zettabytes)
1000^8       Y      yotta (yottabytes)

 

On the other hand, the version we typically refer to are the IEC ones, with JEDEC (SI) prefixes.

 

Multiplier   Abbr   Prefix (Example)
1                   NONE (bytes)
1024         Ki     kibi (kibibytes)
1024^2       Mi     mebi (mebibytes)
1024^3       Gi     gibi (gibibytes)
1024^4       Ti     tebi (tebibytes)
1024^5       Pi     pebi (pebibytes)
1024^6       Ei     exbi (exbibytes)
1024^7       Zi     zebi (zebibytes)
1024^8       Yi     yobi (yobibytes)

 

The error between them can be calculated as 0.9765625^N, which can be multiplied by 100 for the percent, where N is the exponential power (kB/KiB is N = 1, MB/MiB is N = 2, etc.). (I.e. 1kB is actually 97.65625% the size of 1KiB, 1MB is 95.367432% of 1MiB, etc.)

 

When doing these speed tests, I assume CA3LE measures the result as bytes per second divided to the smallest whole measurement, if he's dividing by 1000 then the SI/JEDEC prefix is appropriate, if it's 1024 then the IEC is most accurate, but in today's society the majority of end-users don't actually know the difference.

 

If you want the actual IEC definitions, IEC 60027-2 (parts of which are succeeded by ISO/IEC 80000) has the exact specifications.

 

You can find some information on these prefixes in NIST and Wikipedia, which both reference the IEC 60027-2 standard:

Definitions of the SI units: The binary prefixes - NIST

Binary prefix - Wikipedia

IEC 60027 - Wikipedia

 

When converting to bits, simply multiply whatever number by 8. The byte/bit distinction is made by a capital 'B' for 'bytes', and a small 'b' for 'bits', though as the OP pointed out the small 'b' for 'bits' is generally frowned upon in the lab view, I suggest that most of our users won't really be worried about that particular distinction, but it is still a technical inaccuracy. (So 1KiB = 8Kibit = 8.192kbit = 1.024kB = 1024B.) Thus, when writing speeds they *should* be kbit/s if the measurement is actual bits. If it's bytes, then kB/s or KiB/s depending on the division factor.

 

If anyone wants I can demonstrate several conversion proofs that go into great detail (from the old university Calc 1-3 days) to describe exactly what happens and how, but I feel that's a bit overkill.

 

 

Thanks,

EBrown

Share this post


Link to post
Share on other sites

Thanks nanobot,

 

I agree with a lot of what you wrote.  In this case however, I'm pretty sure the speed being report was meant to be in bits per second not bytes per second.

Assuming that's correct, then the KiB unit wouldn't apply as that's a binary multiple of 8 bit octets.

 

ISO/IEC 80000-1 (Part 1 of ISO/IEC 80000) is the general section that covers quantities and units.  ISO/IEC 80000-13 (Part 13 of 80000) is the part that covers information science.  It has some other great references.

 

I've found the NIST page to be a very good starting out point for folks wanting to get acquainted with units.  BIPM is the international body that regulates/maintains the SI and NIST is the US agency that participates in the international quantities and units standards work for us.  So if one wants to go to the source, the BIPM page can be found here:  http://www.bipm.org/en/measurement-units/

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now