Just saw a comment on IC that my Facebook-trained mouse scurried to 'like' only to be disappointed as usual. There's no 'like' button. Since I never use the 'flag as abusive' button I thought I would use it as a 'like' button. I would simply declare once that a flag from me indicates that I liked the comment and from then on everyone would know the convention. Of course even before you finished the last sentence you realized that this is an idiotic “idea.” For one thing, how would the readers know that the flag is from me? This obvious objection reflects our brains intuitive way of saying, “A simple on/off flag doesn’t have enough information capacity to carry the message.” You can’t fit two gallons of doogh in a one gallon jug! By the same token you can’t fit a two bit message in a one bit flag. Most of us know what doogh is, but what is a “bit”?
One bit of information is the answer to a single “yes/no” question. If you ask a question that can be answered with a single “yes” or “no” then the answer you get carries one bit of information. In the case of the 'flag as abusive' transformation to a 'like' flag' I'm one bit short: Is it flagged (we already have this)? and is it from Ari (We're missing this)? Notice that it doesn’t matter if the answer is “yes” or “no,” ; the bit still gives you information. Why? Because if you ask someone on the phone if she is naked or not and she truthfully says “no,” then you know that she is dressed. A ‘no’ answer to tails is the same as a ‘yes’ answer to heads.
We were pretty smart in quickly figuring out the information capacity of a comment flag, but are we as intuitive when it comes to larger amounts of information. For example take a special game of 20 questions where you have to guess an integer number between one and two million. Does your brain immediately jump up and say, "Olaagh, there’s not enough information in 20 yes/no answers to always nail down a number between one and two million?” I will leave this as a puzzle for the reader to figure out what is the maximum number of manghool or non-manghool objects you can choose from when playing a fair game of 20 questions. Meanwhile on to civil disobedience.
IC members have been asking for a 'like' flag since Mossadegh appeared on the cover of Times. Similarly Iranians have been demanding civil liberties for even longer. In the case of the IC “abusive flag” an act of civil disobedience would be for a large enough number of people to start using it as a 'like' flag thus rendering its information content useless. Readers and admin wouldn’t know what the flag meant anymore. Of course, this would seriously inconvenience IC citizens who use the flag for its original purpose. Also it may annoy bloggers who take more offense at the possibility of a flag meaning “sucks” than can be made up for by the possibility that the flag means ”fantastic” So I am NOT recommending like-button activism to IC members, just illustrating how a single bit can make a difference even in protest movements.
What are some examples of bits outside the world of offensive comments, computers and politics? Well, the Universe is reallly nothing but bits of information. How does the Moon know the Earth is there so that it can revolve around her. If you took the Earth away, the Moon would fly away in a straight line. So the Moon must somehow be getting information from the Earth. What sort of yes/no questions is the moon asking the Earth all the time to keep its orbit? To exist means to interact with whatever else is in the Universe, so existence relies on the bit to…well, exist. Recently there’s been a lot of talk in the media reducing God to a particle found at CERN. I’m not of the Higgs particle religion. Worshipping the higgs belongs to the jaheliyya because the higgs itself relies on the bit for its existence. So the Bit is the True Religion and its Earthly manifestation is the 'flag as abusive' bit. Call me Bitollahi!
Image: Portrait of Professor Fazlollah Reza, leading expert in information theory, by artist Kamran Khavarani.
Recently by Ari Siletz | Comments | Date |
---|---|---|
چرا مصدق آسوده نمی خوابد. | 8 | Aug 17, 2012 |
This blog makes me a plagarist | 2 | Aug 16, 2012 |
Double standards outside the boxing ring | 6 | Aug 12, 2012 |
Person | About | Day |
---|---|---|
نسرین ستوده: زندانی روز | Dec 04 | |
Saeed Malekpour: Prisoner of the day | Lawyer says death sentence suspended | Dec 03 |
Majid Tavakoli: Prisoner of the day | Iterview with mother | Dec 02 |
احسان نراقی: جامعه شناس و نویسنده ۱۳۰۵-۱۳۹۱ | Dec 02 | |
Nasrin Sotoudeh: Prisoner of the day | 46 days on hunger strike | Dec 01 |
Nasrin Sotoudeh: Graffiti | In Barcelona | Nov 30 |
گوهر عشقی: مادر ستار بهشتی | Nov 30 | |
Abdollah Momeni: Prisoner of the day | Activist denied leave and family visits for 1.5 years | Nov 30 |
محمد کلالی: یکی از حمله کنندگان به سفارت ایران در برلین | Nov 29 | |
Habibollah Golparipour: Prisoner of the day | Kurdish Activist on Death Row | Nov 28 |
گروهک فریب خورده و مزدور Bitollahi
maghshooshWed Jul 25, 2012 07:18 PM PDT
The Information Ministry's Information Technology dep't has uncovered further evidence of the mischiefs of the so-called "Bitollahi" deviant cult. This mercenary grouplet has tried in vain to hide the sources of its seditious practices by covering up a bit.
Between the 1st & 2nd letters of its name a bit has been circumcised. The bitten-off bit can assume either of 2 revealing values: either an "e", as in "engelissi", or an "r", as in "rahbari", thereby identifying them as both "Beitollahi" and "Britollahi". There's mounting evidence of further treasonous undertakings, such as taking a whole byte out of their Bits. The people's deputies have demanded swift & decisive action to uproot such bitiful violations of the nation's sanctities.
Bitollahi persecution
by Ari Siletz on Mon Jul 23, 2012 03:22 PM PDTMaghshoosh the "Indian argument" is often used by anti-Bitollahis to misrepresent and underestimate our spirituality. But our ancient writings, some dating back to 2012 AD, assert that it does not matter which we call The One and which The Zero. Swapping the zeros and ones in the Bitollah Bible gives exactly the same divine message. In fact this miracle is solid proof of the existence of the Bit. Try swapping letters in any other so called sacred text and see what happens! This proves that either God does not exist, or if He does He is an inferior manifestation of The Bit. As St. Littlebit said so profoundly, "The one's complement is but an illusion."
Bitollahis are heretics
by maghshoosh on Mon Jul 23, 2012 02:06 PM PDTThe Bitollahi revelation that "In the begining there was the Zero and The Bit said let there be One," is anachronistic & flies in the face of accepted wisdom. The formal concept of positive integers, which begin with 1, has existed tens of millenia earlier than the formal concept of zero. Early humans must have known how to count positive integers beginning with 1. But the formal concept of 0 was introduced by Indians no earlier than the 5th century.
Also from the purely bit-theoretic point of view, 1 & 0 carry the same amount of information, like the heads or tails of a coin; one is not more primordial than the other. When you introduce arithmetic, then 0 & 1 part ways.
Time to Digest
by Zorumbaa on Mon Jul 23, 2012 09:55 AM PDTThanks Ari, now it is my time to digest what you put on the table. À la carte not yet full service butI appreciate additional comments. Frankly a “bit” of what I read is hard to swallow, but I try to wash it down with something! I am really interested in the main course, the all mighty big turkey named god (NO OFFENSE, please! It is by me and for me only!) BTW, have you read The Yellow Dog and his Brother the Jackal? What do you think?
Age old issue
by Ari Siletz on Mon Jul 23, 2012 01:16 AM PDTThanks, Aynak, for taking the time to clarify the discrepancy. I'm sure you know that the bothersome "0th or 1st" question is a also a traditional source of ambiguity for Iranians when it comes to calculating our ages. Some conventions say you are zero years old when you are born and others say you have entered your first year. The former and the latter being the same age yet one year apart!
Zorumbaa, not sure what bringing to a "FULL" means in this case. Do you mean the Moon-Earth issue or the God business? Regarding Moon-Earth , an information theoretic re-interpretation of the laws of physics hasn't fully arrived but seems to have been further incentivized with the recent discovery of the holographic principle (discussed in the Susskind video link). As far as God goes, here is chapter zero, verse zero of the Bitollahi Bible: "In the begining there was the Zero and The Bit said let there be One." Of course the entire book is in binary and we bitollahis each have a different code book...which is how we each maintain a personal relationship with The Bit.
Depends on your interpretation of first iteration :)
by aynak on Sun Jul 22, 2012 11:11 PM PDTMy program starts at 0. So if your picked one number say 1,000,000 my program will report it found it in 0 iteration (not 1). If you take that to be 1 iteration, then you will get 21 for worst case. The theoretical number of needed guess in a binary search for the a number in the range 2**n, would be n. So 4,000,000 would fall in the range 0-2**23=4194304 (i.e 23 guesses my program would say 22) and 8,000,000 would fall in the range
2**24= 8388608 or 24 guesses (mine would say 23). OK don't fire me ;)
BTW, even if I changed my program to start iteration at 1 instead of zero, please note that "max" gusses needing 21 iteration in the range 0-2,000,000 is 951425. So you have a better than 50% chance to still get a number that can be guessed with 20 guesses or less. :)
That is all I got!
by Zorumbaa on Sun Jul 22, 2012 07:03 PM PDTThat is all I got, a referrence to SUPPER source called Wiki. Thanks anyway Ari, you are cool and I am partially satidfied! Can you bring it to a FULL next time?!
Maghshoolsh, Aynak
by Ari Siletz on Sun Jul 22, 2012 03:57 PM PDTYes, theoretically the maximum number of objects in a fair 20 questions game should be 2 ^20 or 1,048,576. Aynak, can you run your program with MAX_VALUE at 4 million and 8 million? Curious to see if the discrepancy with your program is just a matter of one extra count.
20 or 21 bits?
by maghshoosh on Sun Jul 22, 2012 03:11 PM PDTSince 2 to the power of 20 is less than 2million (but more than 1mill.), shouldn't it generally take 21 questions to determine the number between 1 & 2million?
Zorumbaa
by Ari Siletz on Sun Jul 22, 2012 01:02 PM PDTThe gravitational field or for that matter any physical interaction can be viewed as a manifestation of information. Here's a good wiki article on the information approach to physics. This video of a Leonard Susskind physics lecture shows how in practice physicists on the frontier sometimes think in terms of information bits.
you are welcome Ari
by aynak on Sun Jul 22, 2012 01:25 PM PDTAnd you can flag me any time also. Personally I have not been flagged either :)
As far as program, unfortunately the indentation has all been screwed up :( (i.e I never submit a program without proper indentation in real life).
But if you notice, I go through a loop starting from 0 to the MAX_VAL and then it does a find() for each value. (that's why I said it does exahustively go through all the numbers between 0 to MAX. If you look it also keeps the worst case so far. Turns out if the number was 1 (it takes 20 iterations to converge which is the worst case also). Then it went all the way to 2,000,000 and it still found 20 to be the worst case.
Of course you can change the program to do a simple prompt, and scan a specific number, however that would not prove it either way (that 20 is the max), unless you were to input all 2,000,000 values then that would not be smart :) to do the same as what the loop does.. There is actually a mathematical proof for it as well (to see why number 20 here is worst case), but it was quicker to write a program to do all permutations. However if folks are having problem with bits and bytes, then number theory (particularly for Saltanat-Talabs) will likely not help :)
Aynak, thanks for flagging my comments
by Ari Siletz on Sun Jul 22, 2012 12:26 PM PDTAnd many thanks for program. I'm quite rusty on my "C" so let me know if the following critique of it is amateurish: The program does not have an input where the user gets to choose a random 20-questions value for the algorithm to find. It seems "main()" takes MAX_VAL (two milllion) and finds its way to it through a binary search. I would be curious to see the result if you added a random number generator that picks number between one and two million and feeds that value into find(). This would be a more direct confirmation (if confirmed) of Anahid's belief that 20 questions is enough to guess the number [every time]. As for me, I'm betting on the new program being successful only about 52% of the time. But I've lost more bets than I have won, so we'll see.
Some Questions Ari Siletz
by Zorumbaa on Sun Jul 22, 2012 11:23 AM PDTQuote: “So the Moon must somehow be getting information from the Earth. What sort of yes/no questions is the moon asking the Earth all the time to keep its orbit? “Many times I wondered about this type of questions, what is out there that keeps moon in its place and not allowing it to fly it into abyss in a “straight” line? I was taught by some high school teachers of Zamin-Shenaasy that is called gravity. If Earth didn’t like Moon anymore and gave up on her is it possible that she could be picked up let’s say by Jupiter and become Jupiter’s Moon?! Of course my teachers may not know anything about bit and HOW Moon receives directives from Earth, and why she follows?
Quote: “Recently there’s been a lot of talk in the media reducing God to a particle found at CERN.” I really did not know if Higgs was “worship able,” although I guess you and me and everyone else can have our own Bots! Anyway, how can you “REDUCE” nothing to less than NOTHING?
I am waiting for the exact definition of a’ ”Flag” from ADMIN and other experts so if need be do my best to be “politically Correct!”
here you go Ari
by aynak on Sun Jul 22, 2012 11:08 AM PDTQuick summary:
1-I flagged both of your comments. Let's see how dense the Admins are. BTW, this is the first time I have flagged any post on I.C
2-You are not Olagh for thinking 20 questions would give you the answer for a random number between 0 and 2,000,000. In fact in that sample size the longest binary search would result in no more than 20 hits (i.e questions). (conicidentally, Ollaghs are not only not dumb, they are rather smart animals, I am about to do a write up on this matter).
I did a quick program to verify this and is included below. :)
3-You need an execustive summary for your article, most people here (except for Saltant Talabs) understands bits and bytes rather well. And as for Saltant-Talabs, bits and bytes is the last of their worries. I find it rather humorous that D.K should be talking about Orwellian Nightmare:)
Having said all of that, here's my summary for admin:
1-I flagged both of Ari's post based on his request and also that I liked them both :) Ironically, the original post itself, does not have such flag. So one can not express like/dislike for the original idea presented ?!?
2-Nothing is sacred, so please apply that to this out-dated site as well. Even the football forum that I used to frequent a long time ago have had make-overs to have thumbs up/down column with a number associated with it, common how hard is that?
cheers, and good post Ari.
p.s Hope you don't get banned becaue of the two flags ;)
// Binary search for Ari to show a random number picked
// between 0-2000000 can be found with 20 iteration(yes/no) or less.
// (this program exhaustively goes through 0-MAX_VAL and finds the largest
// number of hits (i.e, questions that Ari asks).
#define MAX_VAL (2*1000000)
#define MIN_VAL 0
int debug = 0;
int
find(int i)
{
int h = MAX_VAL;
int m = h/2;
int l = MIN_VAL;
int count = 0;
do {
if (i < m) {
h = m-1;
m = (h+l)/2;
if (debug)
printf ("count %d: l=%d m=%d h=%d\n", count, l,m, h);
} else if (i>m) {
l = m+1;
m = (h+l)/2;
if (debug)
printf ("count %d: l=%d m=%d h=%d\n", count, l,m, h);
} else {
if (debug)
printf ("found Val = %d %d \n", m, count);
return (count);
}
} while(++count);
}
main()
{
int i = 0;
int cur;
int max_sofar = 0;
while (i<= MAX_VAL) {
cur = find(i++);
if (cur > max_sofar) {
max_sofar = cur;
printf ("Val = %d max_search %d \n", i-1, max_sofar);
}
}
printf ("total search %d\n", i);
}
thanks for an interesting blog Ari
by Anahid Hojjati on Sun Jul 22, 2012 09:46 AM PDTtwenty questions are enough to guess the number between 1 and two million.
Thank you Ari Jan
by Darius Kadivar on Sun Jul 22, 2012 01:07 AM PDTI agree ...
Darius
by Ari Siletz on Sun Jul 22, 2012 12:54 AM PDTI read your linked blog; thanks. It made me think of this article by Rem Reider of American Journalism Review who argues that anonymous comments of are bad for business:
"Continuing to allow anonymous sniping hardly seems to be in the
self-interest of news outlets. Sure, everyone is desperately chasing
eyeballs as a way to increase advertising. But rare is the advertiser who would want to be associated with the ugliness of many comment sections."
Reider may be right about the ugliness, but he may be wrong about advertisers having reason to care about association with ugly behavior. The appeal of the anti-hero in fiction seems to extend to real people with fictitious names. Go figure!
My Orwellian Nightmare ...
by Darius Kadivar on Sun Jul 22, 2012 12:02 AM PDTMY ORWELLIAN NIGHTMARE: Flagging Abuse based on Political Correctness is Not Fair
Technical note
by Ari Siletz on Sat Jul 21, 2012 09:41 PM PDTIn the second paragrah it says, "One bit of information is the answer to a single “yes/no” question." To be accurate you have to add "..where 'yes' or 'no' are equally likely." So I've taken some artistic license to quickly get the concept across.