2.2.3 - Non-Canonical SI Prefixes
2.2.3
Non-Canonical SI Prefixes
Go Back Home
Return to 2.2
PREV<< 2.2.2 - Modern Standard SI
INTRODUCTION
The popularity of the SI prefixes has, not surprisingly, lead to a slew of imitators. All kinds of people have suggested to either make certain modifications or additions from amateurs to pranksters to even serious professionals. Some of these amount to hoaxes. The would-be coiner will have a list of SI prefixes including all the canonical entries along with a few additional ones of their own making, and they try to pass them off as canonical through misinformation. These "fake" prefixes then get replicated on other sources and if they multiply enough, they may seem to be part of the canon. At least this would be their hope. Usually these hoaxes are revealed pretty quickly, as they can be checked against official sources. Others make open proposals. These are not considered hoaxes because the author is acknowledging themselves as the source of the proposal. These "proposals" range from ridiculous joke proposals, such as something you might see in a blog, to serious proposals that people write articles about. To date, none of these proposals have been considered for inclusion with the canon.
While an exhaustive account of the history of non-canonical SI prefixes is nearly impossible to construct, we can at least look into some of the major and popular attempts made to "go the extra mile" with the SI prefixes.
OBSOLETES
Perhaps the first prefixes to become non-canonical were Myria- and Myrio-. Recall that these were introduced with the original SI Prefixes in 1795 and later dropped in 1960. Since then they have been defunct, and use of them is technically non-canonical. Myria- is 10^4 and has the unit symbol ma or my , while Myrio- is 10^-4 and has the unit symbol mo .
There is of coarse a whole history of "conventions" that were used and later discarded in the canonical SI system. These prefixes, or in almost all cases, compound prefixes, can be termed "Obsoletes".
Here is a small list of obsolete compound prefixes, and prefixes I have found [1] ...
OBSOLETES :
Compound Prefix Compound Symbol Multiplier Remark
myria- ma 10^4 Instated 1795, deprecated 1960
myrio- mo 10^-4 Instated 1795 , deprecated 1960
decakilo- dak 10^4 1960 use of compounds forbidden
decimilli- dm 10^-4 1960 use of compounds forbidden
hectokilo- hk 10^5 1960 use of compounds forbidden
centimilli- cm 10^-5 1960 use of compounds forbidden
kilomega- kM 10^9 Possible use : 1919 ~ 1960 [2]
millimicro- mµ 10^-9 Possible use : 1919 ~ 1960
megamega- MM 10^12 Possible use : 1919 ~ 1960
micromicro- µµ 10^-12 Possible use : 1919 ~ 1960
The last 4 compound units are obsolete for obvious reasons. Once the new SI prefixes, giga, nano, tera, and pico were introduced, such compounding was no longer neccessary. The first 2 were discarded for infrequent use. The remaining 4 were discarded because of a decision to disregard compound prefixes. Thus things like hectokilo- are no longer used, even though there is no substitute prefix available. Some have even suggested dropping the use of deca, deci, hecto, and centi all together, as they don't fall within the power of 1000 scheme that the rest of the SI prefixes do.
One might ask though, why can't we simply extend the modern SI system the way people used to extend it when they lacked a prefix ? For example Can't we use kiloyotta, or milliyocto ? Officially this is not permissible. Also such units are almost never used, so it's a rather moot point at the moment.
So compounds are ruled out. But what about original prefixes ? Are there other original prefixes besides the canonical SI prefixes ? There are in fact more than is needed. People are fascinated by the esoteric and obscure. Many people have set out to create their own extensions to the SI system. Some of these are quite ambitious and others useless. Let's begin and explore this exotic and difficult to navigate world, the Forbidden Vault of Non-canonical prefixes ...
BOGUS PREFIXES
Before I begin I would like to make note that the information provided here is the result of my own internet research. Although all sorts of "bogus prefixes" exist online, it is very difficult to do serious research on them. On problem is pin pointing exactly where these bogus terms originate. Part of that is do to the hoaxers themselves, and the other part has to do with the nature of the internet. Think of the internet as a room full of mirrors. Information gets replicated over and over until it is difficult to pin point it's source. Often people will mention a term without referencing their source, and thus the trail runs cold. I have made some effort to dig into the origins of these terms. What is offered here is partly speculation based on the sources I've looked at, and should not be taken as absolute fact. I hope it is informative and helps others to do further research in this area.
Note : All bogus prefixes will be written in italics throughout the rest of this article.
THE ALEX LOPEZ-ORTIZ HOAX ?!
I'd like to begin with what seems to be one of the earliest internet hoax of bogus SI prefixes. The prefixes hepa, otta , nea , dea , una , ento , fito , syto , tredo , and revo are typical examples of bogus prefixes. Robert Munafo's popular website on large numbers only has this to say on the subject ...
" All of the "SI prefixes" in italics are unofficial. The following prefixes can be dismissed immediately: hepa, ento, otta, fito, nea, syto, dea, tredo, una and revo; they were all apparently from a popular Internet rumor or hoax. " -- Robert P. Munafo, Large Numbers at MROB [3] .
Despite Robert Munafo's respectable reputation even he does not mention the original source of this "popular hoax".
These same prefixes are also listed on the numericana website on their popular disambiguation table [1]. Here they are all listed under bogus. The first thing to notice is that these terms almost always appear together as a group, as if they all originate from the same source. Another interesting thing is that besides hepa, otta, ento, and fito they all represent denominations not provided for within the official SI system. More important however is the fact that hepa- , ento-, otta-, and fito- can be replaced by the official zetta-, zepto-, yotta-, yocto- respectively. This means that either the hoax occurred before 1991 when the canonical terms were introduced or the hoaxer was unaware of them at the time. Personally I find the first possibility more acceptable.
I attempted to trace back these terms, but no where did I find mention of a "first source". What I did find only deepens the uncertainty. There is one table I found at two links that uses these bogus prefixes , but treats them as part of the canon [4][5]. In each case, the author of the table was someone by the name of Alex Lopez-Ortiz. Apparently he is part of Department of Computer Science at the University of Waterloo. Being that there is not even a hint that these terms are "fake" according to the tables and articles by Alex Lopez-Ortiz, it would seem he is a potential candidate for being the originator of these terms. It's also possible that he inadvertently was lead to believe they were real by yet another source. One interesting note however is that he does not cite any source for his prefixes. So we reach a dead end here.
Another weird thing about this is that the copyright dates range from 1993 to 1996. This conflicts with my theory that the terms were coined prior to 1991. One would have to conclude that if Alex did coin the terms that he must of not known of the new SI proposals. Could he have possibly known and wanted to obscure them on purpose ?! This just seems to lead to more questions and less answers.
In any case the numericana site seems to confirm both my suspicion and doubt in regards to Alex Lopez-Ortiz being the origin of this particular hoax [6]. Although numericana lists him as the originator of the hoax, it puts a question mark next to his name. I have not been able to find out anything more conclusive about the origins of these prefixes.
Although we can't be entirely certain when and where this hoax was first perpetrated, we can at least agree on what these terms mean. That's because they are consistently cited as representing specific denominations [1][4][5] . The table below presents the bogus prefix, followed by a suggested prefix symbol [7] , and lastly the multiplier ...
ALEX L ORTIZ PREFIX PREFIX SYMBOL ?[7] MULTIPLIER
hepa- H 10^21
otta- O 10^24
nea- N 10^27
dea- D 10^30
una- U 10^33
ento- e 10^-21
fito- fi 10^-24
syto- s 10^-27
tredo- t 10^-30
revo- r 10^-33
The derivation of the prefixes is unusual. On the one hand, the large scale prefixes seem to follow a greek pattern, but the small scale prefixes follow no pattern that I can recognize.
We can see that hepa- is alot like the greek "hepte" meaning 7 ( Note that the multiplier 10^21 is the 7th power of 1000 ). Likewise otta- is alot like the greek "okte" meaning 8. nea- is similiar to the greek "ennea" for 9, and dea- similiar to the greek "deka" for 10. It follows that una- should meaning 11, but 11 in greek would be something like "hen-deka", while in latin it would be "undecim". It seems closer to the latin word "una" for 1.
The small scale prefixes are much harder to come up with a rationale for. ento- doesn't sound like anything relating to 7. The names only become more arbitrary as we continue down the line with fito- , syto- , tredo- , and revo- . Only the hoaxer would know what these were derived from.
If you find out anything interesting in your own research feel free to email me so I can make modifications to this article ( my email address is listed at the bottom of my homepage ).
THE MORGAN BURKE PREFIXES
Apparently a hacker by the name of Morgan Burke made a rather whimsical proposal in 1993 for how to extend the SI prefixes. He suggested the use of these additional prefixes : harpi- (10^27) , grouchi- (10^30) , harpo- (10^-27) , and groucho- (10^-30). The proposal met general approval on Usenet [8] .
Several sources confirm that these prefixes were first proposed by Morgan Burke. Robert Munafo makes a brief note of them at the bottom of his "SI prefix" sub-article for the sake of "completeness" [3] .
The prefixes themselves are based on members of the "Marx Brothers" [12] . They were a popular group of comedians composed of 5 brothers who worked in the early days of film. The 5 brothers were each given aliases of sorts, namely : Chico , Harpo , Groucho , Gummo , and Zeppo. Hence the pattern seems to be to take a marx brother name as a small scale prefix ( the lowest power that is not being represented by another prefix ), then form a large scale prefix of the same magnitude by changing the last letter in the name ( always o ) with " i ". Thus if harpo- is 10^-27, then harpi- must be 10^27. Likewise if groucho- is 10^-30 , then grouchi- must be 10^30. This has lead some to suggest that the system could be extended using the missing 3 brothers, namely : Zeppo , Gummo , and Chico [8][9] .
The problem is that there is no conventional way to list these names. One possibility I would suggest is list them by date of birth. If we list them as 1st born, 2nd born , etc. then the order is clearly Chico, Harpo, Groucho, Gummo, and Zeppo. The problem with this is that Morgan Burke has already used Harpo as his first prefix after the canon. Interestingly 2 sources list the proposed name/prefixes in the same order [8][9] . It is tempting to use this as a basis for extension. However neither of these sources actually assigns any kind of value to these potential terms. Therefore we would be wrong to conclude their is an established order, and most likely the order is the result of one source being used to build the other, or just a coincedence. For this reason the possible prefixes Zeppi- , Gummi- , Chici- , Zeppo- , Gummo- , and Chico- are ambiguous. Although we could at least assume that they are all of greater magnitude than grouchi- and groucho-. I leave to others to make sense of these inconsistencies in extension. Here is a table none the less ...
Morgan Burke Prefix Prefix Symbol?[10] Multiplier
harpi- HA 10^27
grouchi- GR 10^30
harpo- ha 10^-27
groucho- gr 10^-30
zeppi- ZE 10^3N where N = {1,2,3,...} & N>8 [11]
gummi- GU 10^3N where N = {1,2,3,...} & N>8
chici- CH 10^3N where N = {1,2,3,...} & N>8
zeppo- ze 10^-3N where N = {1,2,3,...} & N>8
gummo- gu 10^-3N where N = {1,2,3,...} & N>8
chico- ch 10^-3N where N = {1,2,3,...} & N>8
To my knowledge, no one has yet made an official proposal on how to either 1. include these 6 extra terms into Morgan Burkes system, or 2. How Morgan Burkes system should be modified to accomedate the new terms. Personally I like the idea of listing them in order of birth, but this of coarse is Morgan Burkes system and so I think his decision in regards to extension should take priority. To my knowledge Morgan Burke has never attempted to futher extend his system.
THE JEFF K. ARONSON HOAX
Back in 2001-2002 a pharmacologist named Jeff K. Aronson apparently produced hoax tables with the standard SI prefixes and a few of his own devising. This is perhaps one of the first really successful hoaxes, because people still seem to be confused by their apparent normality. Unlike the more pranksterish proposals, the set created by Aronson seem to be right at home amongst the standard prefixes. To extend the standard system Aronson observed that the term zepto- could be formed by combining z with "septo" a variant on the latin word for seven, and yocto- could be formed by combining y with "octo" the latin word for eight. Assuming this pattern continues one could form new prefixes by going backwards through the alphabet , attaching them to the proper latin number, and then changing the last letter to an "o". It can be seen that after milli-, all the small scale SI prefixes end with o, ex. micro- , nano- , pico- , femto- , atto- , etc. Based on these observations Aronson continues after yocto- with xenno- (10^-27) , weko- (10^-30) , vendeko- (10^-33) , and udeko- (10^-36). For the large scale prefixes he uses xenta- (10^27) , wekta- (10^30) , vendeka- (10^33) , and udeka- (10^36). Aronson also created prefix symbols to accompany them. Using the capitals for the large scale prefixes we would have X , W , V , and U, and using the lower cases for the small scale prefixes we would have x , w , v , u. It has been suggested that Aronson did not continue further because of a small snag to this system. The next prefixes would require the use of T and t, but since T is already used for tera- the system can not continue.
I have attempted to trace the original hoax table by Jeff K Aronson, but I have not succeeded. It may be that the page no longer exists. This means that these prefixes are now being maintained solely by those reporting about them, and not by the original hoaxer. Despite the fact that these are "bogus", they do present an almost acceptable kind of proposal. For reference , here is another table ...
JEFF K. ARONSON PREFIXES[13] ARONSON SYMBOL MULTIPLIER
xenta- X 10^27
wekta- W 10^30
vendeka- V 10^33
udeka- U 10^36
xenno- x 10^-27
weko- w 10^-30
vendeko- v 10^-33
udeko- u 10^-36
The popularity of Aronson's system eventually lead one other to follow his lead with yet another proposal that has become one of the most popular extensions on the web.
THE JIM BLOWERS PROPOSAL
Prior to 2009 but after the 2002 Jeff K. Aronson Hoax a guy by the name of Jim Blowers took Aronson's basic idea an took it one step beyond. Unlike Aronson , Jim Blowers announces immediately that he wishes to extend the system, thus this is an informal proposal NOT a hoax. A table of his extended prefixes is still on his website [14] . This table is often cited from all sorts of different sources. This shows both the popularities of such extensions, and this one in particular.
So what exactly did Blowers propose ? Like Aronson he constructs new terms by going back through the alphabet. However he also alternates between latin and greek roots. The odd terms are latin based while the evens are greek. A side by side comparison of the etymology and the proposed prefixes though reveals that Blowers, like Aronson takes a fair amount of liberty to make them fit. The table shown below makes this clear ...
Note : The prefix symbols were created by Jim Blowers himself to accompany his prefixes [14] .
JIM BLOWER PREFIXES
PREFIX SYMBOL MULTIPLIER ETYMOLOGY?
xona- X 10^27 x - noni(latin)
xonto- x 10^-27 x - noni(latin) - to
weka- W 10^30 w - deka(greek)
wekto- w 10^-30 w - deka(greek) - to
vunda- V 10^33 v - undecim(latin)
vunkto- v 10^-33 v - undecim(latin) - kto?
uda- U 10^36 u - duodeka(greek)
unto- u 10^-36 u - duodeka(greek) - nto?
treda- TD 10^39 t - tredecim(latin)
trekto- td 10^-39 t - tredecim(latin) - kto?
sorta- S 10^42 s - tesseradeka(greek)?
sotro- s 10^-42 s - tesseradeka(greek)?- tro?
rinta- R 10^45 r - quindecim(latin)
rimto- r 10^-45 r - quindecim(latin) - mto?
quexa- Q 10^48 q - hexedeka(greek)
quekto- q 10^-48 q - hexedeka(greek) - kto?
pepta- PP 10^51 p - septendecim(latin)
pekro- pk 10^-51 p - septendecim(latin) - kro?
ocha- O 10^54 o - oktodeka(greek)
otro- o 10^-54 o - oktodeka(greek) - tro?
nena- N 10^57 n - novemdecim(latin)
nekto- nk 10^-57 n - novemdecim(latin) - kto?
minga- MI 10^60 m - icosi(greek)
mikto- mi 10^-60 m - icosi(greek) - kto?
luma- L 10^63 l - unviginti(latin)
lunto- l 10^-63 l - unviginti(latin) - to
You might ask why Blowers chose to stop here ? As he explains he wanted to have as many prefixes as there are official -illions. The -illions as you may recall being mentioned in the previous chapter officially end at vigintillion. A luma- is a factor of 10^63 , that is one vigintillion.
It's pretty clear why this system has been so successful. First of all, it goes alot further than the other proposals/hoaxes. One important thing that leads it a certain authority is that it is a collection of prefixes, a prefix system. This gives them order and context. It also forms fairly acceptable prefix names that don't sound much more riddiculous than the "real" ones. Another important feature is that most of these prefixes are fairly short, usually no more than 5 letters. This helps them blend in with the canonical SI prefixes. Their derivations are alittle bit of a stretch sometimes, but at least they are based on some kind of pattern. In many ways this is one of the best non-canonical systems that has yet been proposed. You almost feel it should be ratified for being so ambitious if nothing else.
But some people may still wonder, why stop at L ? can't we continue further down the alphabet and form even more prefixes !? Why not get all the way to the beginning of the alphabet and be done with it ! Has anyone attempted this half-mad project ? As a matter of fact I did find one rather obscure tid-bit confirming just this ...
THE LONG LOST H.PAUL SHUCH APPROACH
Apparently someone came up with the whole "backwards through the alphabet" idea long before Aronson and Blowers got to it. I discovered there was some named H.Paul Shuch who made mention of just such a system in a letter entitled "The large and the small of it" submitted to Science News 143(14):211 dated April 3 , 1993 [15][16]. This is around the time that Alex L Ortiz was publishing his hoax, and Morgan Burke was making his marx brothers proposal.
What should first be noted is that Shuch's proposal is qualitatively different from all of the above proposals. It is not merely a exhaustive list of prefixes, rather it's a rule to construct such a list, which conserves alot more space.
The idea is simple. We continue after yocto- by going back through the alphabet and using the root "-octo". when the letter is a vowel we can drop the first "o" in the root. Thus we continue with xocto- (10^-27) , wocto- ( 10^-30) , vocto- (10^-33) , ... all the way through to acto- (10^-96). Similarly we continue after yotta- by going backwards through the alphabet and using the root "-otta", again we can drop the "o" when the first letter is a vowel. Thus we continue with xotta- (10^27) , wotta- (10^30) , votta- (10^33) , ... all the way through to atta- (10^96).
For the sake of completeness here is a compact listing of the prefixes (followed by their respective exponents) that would be formed in this manner ...
SHUCH PREFIXES :
xotta 27, wotta 30, votta 33, utta 36, totta 39, sotta 42, rotta 45, qotta 48, potta 51, otta 54, notta 57, motta 60, lotta 63, kotta 66, jotta 69, itta 72, hotta 75, gotta 78, fotta 81, etta 84, dotta 87, cotta 90, botta 93, atta 96
xocto -27, wocto -30, vocto -33, ucto -36, tocto -39, socto -42, rocto -45, qocto -48, pocto -51, octo -54, nocto -57, mocto -60, locto -63, kocto -66, jocto -69, icto -72, hocto -75, gocto -78, focto -81, ecto -84, docto -87, cocto -90, bocto -93, acto -96
This system goes even further than Blowers does. But it does have some problems. Firstly it doesn't provide any prefix symbols. Etymologically it's also rather absurd. Recall that the roots "octo" and "otta" are derived from the latin word for 8. This clearly makes little sense. Furthermore there is even a shuch prefix called "octo" for 10^-54 . It would seem a bad idea to use a latin root by itself as a prefix of any kind. Afterall how are we suppose to interpret an "octometer" ?! It's suppose to be 10^-54 meters, but it sounds more like it should mean 8 meters . Also, how do we distinguish between kotta and cotta when they are being spoken? And the difference would matter because they're 24 orders of magnitude apart ! Some of them also sound alittle risque ( do I need to point them out ? ). Lastly the construction seems too simple to be taken seriously. When compared to the much more rich and varied Blowers prefixes it seems kind of bland in comparison.
Still, it's interesting that these never caught on much. There are plenty of other systems that have fooled people into believing they're part of the canon.
SUGGESTION BY GERARD MICHON
Seems everyone wants a chance to extend the standard prefixes. Gerard Michon is one of the authors of the numericana website. He is responsible for the disambiguation table. After tracking several bogus prefixes he makes a proposal himself [6]. Michon suggests the use of nova- for 10^27 and novo- for 10^-27. He even recommends using the prefix symbols "N-" and "n-" respectively. He says the etymology is based off the french word for 9 (neuf), although "novem" is also latin for 9 and this seems a much closer fit.
Michon Prefix Michon Symbol Multiplier Etymology
nova- N- 10^27 neuf(french) or novem(latin)?
novo- n-* 10^-27 neuf(french) or novem(latin)?
* n- is already being used for nano- so another symbol would need to be used if the Michon prefixes were appended to the BIPM prefixes.
Michon's proposal is not that different than those of Aronson and Blowers. Although it doesn't extrapolate the "backwards through the alphabet" pattern, it does exploit the idea of adapting numbers into prefixes. In fact Michon's system can easily be extended by developing a way to adapt either french or latin numbers.
PROJECTIONS BY ROBERT P. MUNAFO
There is one more system that I'd like to discuss. Robert P. Munafo's popular large number pages have been a reliable and unified source on large numbers for many people for several years ( including myself). I first discover the site back in 2004 when doing my own research into the subject of large numbers. He also has a sub-article which discusses the SI prefixes.
The page url is :
http://www.mrob.com/pub/math/largenum-2.html
The sub-article is the 3rd heading down the list ( SI Prefixes ). In the sub-article Robert gives a brief history of the standard prefixes and then discusses briefly some of the more popular bogus prefixes out there.
At one point Robert speculates about what kind of prefixes would truly be included in the canon. He makes up a set of projected or hypothetical prefixes that might one day be accepted. Here is the original quote from the same sub-article :
" If BIPM decides to adopt further prefixes for 1027 and 1030 and their reciprocals 10-27 and 10-30, they will probably adopt something vaguely resembling names for nine and ten for a similar reason — perhaps something like novetta, novemo, decetta and decemo. If so they would almost certainly be assigned two-letter abbreviations such as "No-", "De-", "no-" and "de-" because N-, n- and d- are already used for other prefixes. "
--Robert P. Munafo
I have emailed Robert Munafo to ask about this interesting suggestion he made. His stand is that prefixes should only be adopted as the need for them arrises. This seems to me a completely sensible and level headed stance. Unlike the hoaxers and proposers I have mentioned above, Robert is not making a suggestion, it is pure speculation as to what the BIPM might adopt in the future.
None the less, I feel that these hypothetical prefixes amount to an unintentional proposal. The prefixes novetta- , novemo- , decetta- , and decemo- differ considerably from all the previous proposals and seem relatively tame in comparison. Unlike the Shuch, Aronson, and Blowers proposals, Munafo doesn't consider continuation backwards through the alphabet a neccessary component of extending the SI prefixes. His prefixes are solely based on derivations of the latin words for 9 and 10 ( novem, and decim respectively ).
The advantage of this is clear. If the BIPM did adopt a similiar type of system, it would establish a convienent precedent that could then readily be extended when needed.
Likewise Munafo's system is readily extendable by continuing the trend established with his first 2 prefixes by adapting the latin numerals into new prefixes. He never makes any attempt however to extend it, but the implication is clearly there. This issue will be taken up in greater detail in the next article. Here I only want to discuss the system as Munafo explicitly states it.
The Prefix symbols that Munafo uses for his hypothetical prefixes are new. They were not listed there originally. However Munafo wrote to me in response to my email that he had decided to update his SI prefix article because he noticed it was out of date. The prefix symbols were therefore added sometime on either 2009.03.27 or 2009.03.28.
One interesting thing to note is that if the BIPM does eventually adopt more and more prefixes, there will come a point at which they would have to accept 2 letter prefixes because the 1 letter prefixes would eventually get used up.
In any case we now have yet another possible system to extend the SI system. Here is a table for Reference ...
Speculative Extension of Robert P. Munafo
Speculative Prefix Prefix Symbol Multiplier Etymology
novetta- No- 10^27 novem(latin=9) + etta
novemo- no- 10^-27 novem(latin=9) + emo
decetta- De- 10^30 decim(latin=10) + etta
decemo- de- 10^-30 decim(latin=10) + emo
When and if the BIPM will ever adopt new prefixes, and what they will turn out to be is still up in the air. We will simply have to wait and see if Munafo's prediction proves accurate. In the meantime however I think that these prefixes can be regarded at one of the better theoretical extensions, and certainly the most sane.
THE STRAGGLERS : MISCELLANEOUS JARGON
The 7 systems I have just discussed, the Alex L , Morgan Burke, Jeff K Aronson, Jim Blowers, Shuch , Michon, and Robert P. Munafo prefixes are the only fully worked out systems that I have found. So have we exhausted the list of bogus prefixes ?! of coarse not. There are also plenty of stragglers, prefixes that aren't part of any extended system. In the worse cases some of these stragglers aren't even determinate denominations and get repeatedly misreported resulting in multiple citings with different values !
Brontobytes and other ridiculous blogger babble
Ever heard of a brontobyte ? No it's doesn't exist of coarse, but you can find it at many different places on the web. Typing brontobyte in a google search engine brings up about 9000 hits currently ( as of 2009.03.29 ).
Even wikipedia mentions a brontobyte [17]. In this case however they rightly ascert that it is an unofficial prefix. The article reports that the brontobyte has been reported at several different values on the internet including 10^15, 10^21 , 10^24, and 10^27. The reason for this is probably do to uninformed individuals assuming that a brontobyte must be the prefix after the highest prefix commonly known, and then reporting it has such a value. For example, a few years back few people would have known what a petabyte was. It is possible that the brontobyte was originally a hoax suppose to stand for the term after a terabyte. Evidence to support such a theory is limited however [18] . In any case , a brontobyte seems to be a kind of informal term for the next value above the highest currently accepted value.
If this theory is correct, one would assume that brontobyte would be cited mostly as 10^27 in our time, since a yottabyte (10^24) is the highest offically recognized term. Although it is difficult to check this, there seems to be a large consensus amongst the people who inadverently (or deliberately) use the term that a brontobyte stands for 1024 yottabytes, or approximately 10^27 bytes.
The etymology of bronto- is that it is greek for "thunder". In this sense bronto- actually fits in with the other officially recognized prefixes. After all mega-, giga-, and tera- are all derived from the greek words for "big", "giant", and "monster" respectively.
Among the uninformed, the brontobyte has become kind of the quasi-standard for 1024 YB. This has lead to further hoaxes to extend the silly terms even further. For example one source which lists brontobyte along with the official terms, also includes one more called a "geopbyte" which is supposedly equal to 1024 brontobytes ( ~ 10^30 bytes ) [19]. "geopbyte" bring up 1320 hits on google, so it's less wide spread than brontobytes, but still "on the map" so to speak. There is even a "yahoo answers!" blog where one user cites a geopbyte as if it is part of common parlance [20].
In fact, a geopbyte seems to be "common" enough that one person has already decided to go one step beyond. A user by the name of Richard Hahn jokingly defined the "MacMillanbyte" to be 1024 geopbytes [21]. The post is dated 2008.10.06, so this post has only been around for a few months ( as of 2009.03.29 ). A google search for "MacMillanbyte" currently only recieves 1 hit, and brings up the page with the original post. Thus it is here that we reach the end of the trail. It's is doubtful that this term will successfully spread.
The brontobyte also branches off into a 2nd series of names. Some forums list 2 additional terms after the brontobyte unlike the ones just discussed. A few sources define a "nisababyte" as 1024 brontobytes, and a "zotzabyte" as 1024 nisababytes [22][23]. "nisababyte" bring up 205 hits on google (2009.03.30), and "zotzabyte" brings up 449 hits on google (2009.03.30). This suggests that the geopbyte is the more established continuation after the brontobyte. It should be noted that in one of my sources [22] the moderator Chieftess as a disclaimer at the bottom of her post that says she is not sure about nisababytes, and zotzabytes, because she originally found them on a message board where there were usernames "nisaba" and "zotz". This is likely the origin of the rumor. Unfortunately I have not found that fateful message board.
There is one last twist in the weird story of the brontobyte. Apparently nisababyte and zotzabyte are already well known enough that someone somewhere came up with another name after even this ! One source I found defines a "omegabyte" as 1024 zotzabytes [24] . However there is a disclaimer at the bottom holding that nisababytes, zotzabytes, and omegabytes are only rumors. Ironically the brontobyte is included as a legitimate prefix ! Although omegabyte gets about 7000 hits on google (2009.03.30) these hits usually involve the names of various corporations, and have nothing to do with the rogue prefix. Again we hit a dead end.
Basically the brontobyte breaks off into two branches of terms, one alittle more established than the other. The Table below displays this ...
Bronto- Family of Prefixes
Prefix , Symbol (hits) Commonly accepted value
bronto- , B- (9360) 10^27 or 2^90
geop- , ?- (1310) nisaba- , ?- (205) 10^30 or 2^100
macmillan- , ?- (1) zotza- , ?- (448) 10^33 or 2^110
omega- , ?- (7130)* 10^36 or 2^120
* no relevant results , only 1 known source.
Keep in mind that this table is not static. There will likely be more hoaxes as time goes on making it even more difficult to keep track of all the unusual prefixes floating out there on the internet. Another interesting phenomenon about the internet is that you can't simply hold a mirror to it without affecting it somehow. Most likely the information I have provided here will not make these terms vanish just because I'm "discrediting them as hoaxes" but rather lead to further proliferation of these terms and in turn further hoaxes. This is simply the nature of the internet. The key is not to be taken in by a single source. If someone mentions a prefix, check it.
Non sequiturs
We're really getting down to the fringe now. All the really good systems have already been covered, and even the rather dubious bronto- family of prefixes has been covered. What is left are essentially pranks, or inside jokes.
Here is a good example. Wikipedia is a popular source of quick and relatively reliable information, but it is also a breeding ground for all sorts of corrupt memes desperate for a chance of replication. There is a wikipedia article entitled "FLOPS", which discusses it's use as a computer term [25] . A table on the left discusses various computer speeds such as "megaflops" (10^6 operations ), "gigaflops" (10^9 operations) , and "teraflops" (10^12 operations). All of these are of coarse legimate terms formed from the standard BIPM prefixes. If you go back through the history of this article however you find that this page has been a frequent victim of vandalism. In fact I discovered a rather assuming prank that someone pulled.
One version of the article dated 2008.06.26 lists among the usual terms a "gonnaflop" and defines it as 10^48 operations per second. An amusing play on words, the revision has no known author but only an IP address [26]. Seven minutes later the term is promptly removed in the next revision by a wikipedia user named Rilak, but with this telling footnote left by Rilak about the revision ...
"undid revision 221837852 by [IP ADDRESS OMITTED BY SBIIS] revert rather amusing vandalism " -- Rilak
Amusing perhaps, but extremely short lived. This meme only lived a mere 7 minutes before being terminated ! But of coarse the meme didn't really die because it was stored in the history log within the backrooms of wikipedia where it waited to be replicated by some misguided individual (ie. myself ), and as you can see it has succeeded in creating at least 1 additional copy of itself.
This sort of thing goes on all the time. Most of these short lived pranks never last and are forgotten. There maybe thousands of bogus prefixes of this variety, but they are so poor at replicating that they quickly get swept up and forgotten. It is also not that easy to find these kinds of things. I had to comb the history of the "FLOPS" page to discover the gonnaflop.
Other "fringe prefixes" ( they don't even qualify as bogus anymore because no one debating about whether they're legimate or not ) are simply little insider "jokes". For example Gerard Michon goes ahead and makes up a completely absurd prefix called a "millikan-" with the prefix symbol "mkan-" [6] . As Michon explains it , the prefix is meant to be an abbreviation for milli-kilo-atto-nano- which would be (10^-3)(10^3)(10^-18)(10^-9) which simplifies to 10^-27 being the definition for millikan-. He goes on to claim that it should be in honor of a Robert A. Millikan (1868-1953) to be fair since he was cursed with a name that "makes it otherwise impossible to achieve the same kind of SI immortality as Pascal, Newton, Ampère, Tesla, Weber, or Alexander Gram Bell " -- (Gerard Michon). All of this is just a silly "science joke". Again these kinds of spoof proposals are probably all over the place, but almost impossible to find.
Spoof Prefix Spoof Symbol Multiplier Etymology
gonna- ?- 10^48 "I'm gonna flip !" context : used in gonnaflop
millikan- mkan- 10^-27 abbreviation for milli-kilo-atto-nano-
There are doubtless many other examples of these strange fringe prefixes, which not only are ignored by the BIPM, but by everyone else in fact ! Yet they exist deep in the cracks and crevices of the internet. Hidden deep within late night forums where people minds tend to get alittle awry, or short lived revisions burried deep within the histories of wiki's across the web. If searching for these terms is like a treasure hunt, finding the treasure is the equilavant of finding out that the chest only contained a few old rusty coppers.
The xeraflop scandal
Alright maybe scandal is a bit strong. Only one person seems to be outraged in this case, but the xeraflop illustrates a more serious problem with information on the internet.
Believe it or not, the term xeraflop appeared on an internet article dated 2008.06.09 for the new york times [27] ! The article entitled "Military Supercomputer Sets Record" talks about the militaries new petaflop machine. It states that the petaflop has long been considered a milestone by military and scientific organizations in the united states for computer technology. The article ends with this statement :
"The next thousandfold goal is the exaflop, which is a quintillion calculations per second, followed by the zettaflop, the yottaflop and the xeraflop. " -- New York Times Online article
In one way this is increadible. Here is a completely illegimate term resting comfortably in a major news organizations online article ! And it get's worse. This last line from the New Time article got replicated verbatim (along with the bogus term xeraflop) by several other less reputable news organizations across the web, such as the neoseeker for example [28] .
How did this happen ? And where did this xeraflop term originate anyway ?!
I have done some research and I am almost certain that the term originated in none other place than wikipedia itself. Remember the "FLOPS" page I mentioned which contained the "gonnaflops" term ? Well if you go to the flops page [25] and click on the discussion tab you can find that item 25 on the content list is titled "Xera is a fake prefix" [29][31] .
What we find at this sub-heading is essentially a debate between 2 wikipedia members, Lordvoltron and ExNihilo. What they are debating is the addition or removal of the term xeraflop to wikipedia. Lordvoltron argues that the term should be recognized because various sources make use of it, and he provides 2 links as proof. ExNihilo however has a strong counter argument. He points out that both sources that Lordvoltron cites were written on 2008.06.09 and both are quoting the same text about the roadrunner supercomputer. Furthermore that this same text was then copied many times verbatim across the web. He then argues (although he can not prove directly) that chances are all the sources would eventually lead back to the very wikipedia article they're discussing ( the FLOPS page). Circular reason ? Yes, but it's not ExNihilo who's being circular here.
Reading this argument began to make me suspicious of Lordvoltron. Why was he so concerned with supporting the term xeraflops, and so quick to use the rapid proliferation of the term as justification for the term itself.
What is interesting is that ExNihilo mentions that the origin of the term is concurrent with the date of the articles (2008.06.09). Based on a hunch, I decided to go back through the history of the FLOPS page to a time around this date, and I found that in fact the term "xeraflops" was repeatedly added and removed around this time period.
The earliest occurrance of the term I found was a revision of the flops page dated may 10th of 2008 [30]. And when I looked at the name of the editor responsible for the revision my suspicions were confirmed. None other than Lordvoltron was responsible for the earliest known revision involving "xeraflop". For me this is compelling evidence that Lordvoltron is the source of the hoax, and it also give him a strong motive to defend the term.
Although I can not prove this , it is tempting to speculate. If one wanted to legitimize a term, what would be the easiest way to do it ? First insert the term in a wikipedia article. Then defend it long enough ( it may be deleted but you can simply reinsert it ) for it to be copied by a source outside wikipedia. Let it spread, and then use it's apparent commonality to justify it's use on wikipedia as part of common parlance. Finally let this viscious circle repeat itself until the term becomes a pseudo-standard.
Of coarse even this apparently successful hoax hasn't been terribly successful. A search for the term "xeraflop" brings up alot of irrelevant results. Hoaxes seem to be measured in the thousands of hits, but they rarely make it into the millions.
If you don't believe me consider the following table ...
PREFIX STANDARD VALUE HITS ON GOOGLE
kilo-(BIPM cert.) 10^3 35,500,000 (2009.03.30)
mega-(BIPM cert.) 10^6 141,000,000 (2009.03.30)
giga-(BIPM cert.) 10^9 21,100,000 (2009.03.30)
tera-(BIPM cert.) 10^12 22,300,000 (2009.03.30)
peta-(BIPM cert.) 10^15 30,600,000 (2009.03.30)
exa-(BIPM cert.) 10^18 6,350,000 (2009.03.30)
zetta-(BIPM cert.) 10^21 1,190,000 (2009.03.30)
yotta-(BIPM cert.) 10^24 168,000 (2009.03.30)
xonabyte(bogus) 10^27 240 (2009.03.30)
wekabyte(bogus) 10^30 371 (2009.03.30)
brontobyte(bogus) 10^27 9360 (2009.03.30)
geopbyte(bogus) 10^30 1310 (2009.03.30)
nisababyte(bogus) 10^30 205 (2009.03.30)
macmillanbyte(bogus) 10^33 1 (2009.03.30)
zotzabyte(bogus) 10^33 448 (2009.03.30)
omegabyte(bogus) 10^36 1? (2009.03.30)
xeraflop(bogus) 10^27 543 (2009.03.30)
By these measurements, the brontobyte may be the greatest internet hoax of the present. And unlike the xeraflop I haven't been able to trace the origins of brontobyte.
HOAX PREFIX APPARENT VALUE HOAX Origin
xera- 10^27 wikipedia member : Lordvoltron (2008) ?
I'll close with a statement made by ExNihilo ...
" [The Xeraflop scandal] is a textbook example of how Wikipedia can spread false information very quickly when people consider it authoritative and nobody checks references "
-- ExNihilo , Wikipedia member
Lastly while I see nothing wrong with people inventing new prefixes, I do see a problem with these same prefixes being passed off as legitimate alongside the BIPM prefixes. It's simply a matter of proprietory rights. It is up to the BIPM to decide what it does and does not regard as legimate. To try and slip in terms is equivalent to suggesting that the BIPM officially endorses them. The bottom line is that credit should only be given to the organization and individuals responsible for the origin of any terminology. While this is meant to protect the right of the originators ( not allowing others to take credit for someone elses work ) it should also be understood that it is meant to protect a organizations or individuals integrity (ie. only that organization or individual should be taken as an authoritive source on the nature of the work it performs ). This prevents others from tarnishing another person or organizations reputation.
This is all that I could find on the web in regards to non-canonical SI prefixes. The next article takes all of this one step beyond. In the next article I offer several suggestions and proposals of my own.
NEXT>> 2.2.4 - Sbiis Saibian's Extended Prefixes (2009)
> 2.2
> Home
Source Material & Footnotes :
The Alex L Ortiz Prefixes
[1] http://home.att.net/~numericana/answer/units.htm#prefix : Popular disambiguation table for SI prefixes. It distinguishes between the canonical, the obsolete, and the "bogus". Part of the numericana website. Table includes the "Alex Lopez" prefixes.
[2] Based on my interpretation of the table at the link in citation 1 , it seems that mega and micro were in use before 1960, and unofficially used as far back as 1919. So terms such as kilomega, and millimicro would have been used after 1919 but before 1960 when units like giga and nano were introduced.
[3] http://www.mrob.com/pub/math/largenum-2.html : The quote can be found under the sub-heading "SI Prefixes" on this page. This is part of Robert Munafo's large numbers site.
[4] http://www.cs.uwaterloo.ca/~alopez-o/math-faq/mathtext/node25.html : This rather plain document entitled "Names of Large Numbers" includes the 10 bogus prefixes along with the standard prefixes without any kind of disclaimer. At the bottom we can find that this article is authored by Alex Lopez-Ortiz, and seems to be dated Feb 20 , 1998. Possible source of "one of the earliest internet hoaxes".
[5] http://74.125.93.104/search?q=cache:ngdoT5z8aZkJ:www.cs.uwaterloo.ca/~alopez-o/math-faq/math-faq.pdf+%22Alex+Lopez-Ortiz%22+,+ento+,+fito&cd=3&hl=en&ct=clnk&gl=us : This rather long url will take you to an html version of a pdf document called "frequently asked questions in mathematics" authored by Alex Lopez-Ortiz. On page 44 of this document we can see the same 10 bogus prefixes listed in exactly the same table appearing on the link at citation 4.
[6] http://home.att.net/~numericana/answer/humor.htm#hoax : The table that should appear shows various bogus prefixes along with their known or suspected hoaxers and approximate dates. On this table nea- and syto- are credited to Alex Lopez-Ortiz, and the hoax is dated around 1996-1998.
[7] In all sources of the "Alex L Ortiz Prefixes" there is no mention of prefix symbols [1][4][5][6] . The prefix symbols used here are therefore of my own devising. Feel free to use, modify, or devise your own. They are meant to supplement the standard prefix symbols, except that zetta- , zepto- , yotta- , and yocto- should be removed.
The Morgan Burke Prefixes
[8] http://www.mondofacto.com/facts/dictionary?yotta- : The mention of Usenet was found in this article. Also the article mentions the possibility of future expansion of the Morgan Burke Prefixes.
[9] http://python.net/~eddy/study/value/quantity.py : This looks like it was intended as an html file. Note only does it include the Morgan Burke Prefixes along with the usual ones, but it has a disclaimer explaining that they were first proposed by Morgan Burke although they are "probably unratified". Interestingly it also mentions the possibility of future expansion, and lists the names in the exact same order as the mondofacto article.
[10] The "Morgan Burke Prefixes" have not been previously assigned prefix symbols. The prefix symbols used here are therefore of my own devising. Feel free to use, modify, or devise your own. They are meant to supplement the standard prefix symbols.
[11] These statements should not be taken as "definitions" for these terms, but rather it is a range possible values that it might potentially be defined as by some future hoaxer, or coiner.
[12] http://en.wikipedia.org/wiki/Marx_Brothers : Wikipedia article on the marx brothers.
The Jeff K. Aronson Hoax
[13] The url listed at citation 3 also includes the spellings for the Aronson prefixes I used. Unfortunately the spellings used at the url at citation 6 differ. I am going to trust however that Robert Munafo's source is older and therefore a more accurate representation of the original hoax.
The Jim Blowers Proposal
[14] http://jimvb.home.mindspring.com/unitsystem.htm : This is the famous Jim Blowers table. Because it includes a foreword explaining his intentions and naming scheme it constitutes a legimate proposal. It has also met with fairly good approval.
The Long Lost H.Paul Shuch Approach
[15] http://www.arthistoryclub.com/art_history/Nea : On the bottom of this short stub page is the mention of the H.Paul Shuch Approach. Be warned however, this page has several pop-ups, though they don't seem to cause any serious problem.
[16] http://enc.slider.com/Enc/Weko : This stub found at a completely different url has almost exactly the same information and also makes a brief note of the H.Paul Shuch Approach. This page doesn't have pop-ups however.
The Stragglers : Miscellaneous Jargon
[17] http://en.wikipedia.org/wiki/Non-SI_unit_prefixes : Small stub article on Wikipedia about Non-SI prefixes. Includes a brief mention of the brontobyte along with various cited uses on the web.
[18] http://openc.k12.or.us/sitedocs/styleguide/abbreviations.html : This article reports that BB (brontobyte) used to be 10^15, but has no been replaced by the petabyte. This suggests that the BB hoax goes a number of years back when terms like petabyte were not well known, and then BB was later "replaced" when the next SI prefix became more widely known.
[19] http://www.whatsabyte.com/ : A brontobyte and a geopbyte are defined as 2^90 & 2^100 bytes respectively.
[20] http://answers.yahoo.com/question/index?qid=20080323165318AAQxFkw : The user Bill on this Yahoo answers! blog seems to treat the geopbyte as if it's a real term.
[21] http://www.digitalimagecafe.com/forum/forum_post.asp?id=6774&page=1 : MacMillanbytes coined by Richard Hahn and defined as "1024 geopbytes" on post dated 2008.10.06.
[22] http://forums.civfanatics.com/showthread.php?t=85287 : In the first post on this forum, moderator "Chieftess" lists Nisababyte, and Zotzabyte after fabled brontobyte. Note however that while brontobyte does not have a question mark, Nisababyte and zotzabyte do. Does this mean people think brontobyte is more legitimate somehow ?
[23] http://www.webmasterworld.com/foo/3448874.htm : A second source which lists Nisababyte and Zotzabyte following after brontobyte.
[24] http://www.halobabies.net/forums/viewtopic.php?f=13&t=9743&view=previous : Source of the "omegabyte" defined as 1024 zotzabytes.
[25] http://en.wikipedia.org/wiki/FLOPS : The "FLOPS" article on wikipedia. There have been at least 2 bogus prefixes that have their origins here, and one of them, the xera- actually succeeded in spreading across the net unchecked !
[26] http://en.wikipedia.org/w/index.php?title=FLOPS&diff=221837852&oldid=221836076 : Earlier version of "FLOPS" article dated 2008.06.26. It is the only occurrence of the spoof term "gonnaflops" for 10^48 operations per second.
The Xeraflop Scandal
[27] http://www.nytimes.com/2008/06/09/technology/09petaflops.html?_r=1&oref=slogin : This online new york times article actually mentions the term "xeraflop" ! Ironically the article doesn't even cite it's source.
[28] http://www.neoseeker.com/news/8159-military-supercomputer-reaches-new-milestone/ : Another online article from the "neoseeker" which copies some of the lines from the new york times article verbatim and includes the bogus term "xeraflop"
[29] http://en.wikipedia.org/wiki/Talk:FLOPS#Xera_is_a_fake_prefix : This argument between Lordvoltron and ExNihilo holds clues to the origin of the term "xeraflop".
[30] http://en.wikipedia.org/w/index.php?title=FLOPS&diff=211432628&oldid=211202535 : The earliest version of the FLOPS page I found that included the "xeraflops" term. The revision was made by Lordvoltron on may 10 , 2008. This is convincing evidence that Lordvoltron is the original hoaxer and origin of the xeraflop term.
[31] http://en.wikipedia.org/wiki/index.html?curid=43574#xera : A later battle between Lordvoltron and ExNihilo, again about the xeraflop issue.