Some recent results in mathematics related to data transmission:

1
Some recent results in mathematics related to
data transmission

• Michel Waldschmidt
• Université P. et M. Curie - Paris VI
• Centre International de Mathématiques Pures et
  Appliquées - CIMPA

India, October-November 2007
http//www.math.jussieu.fr/miw/
2
French Science Today
India October- November 2007
Some recent results in mathematics related to
data transmission
Starting with card tricks, we show how
mathematical tools are used to detect and to
correct errors occuring in the transmission of
data. These so-called "error-detecting codes"
and "error-correcting codes" enable
identification and correction of the errors
caused by noise or other impairments during
transmission from the transmitter to the
receiver. They are used in compact disks to
correct errors caused by scratches, in satellite
broadcasting, in digital money transfers, in
telephone connexions, they are useful for
improving the reliability of data storage media
as well as to correct errors cause when a hard
drive fails. The National Aeronautics and Space
Administration (NASA) has used many different
error-correcting codes for deep-space
telecommunications and orbital missions.
http//www.math.jussieu.fr/miw/
3
French Science Today
India November 2007
Some recent results in mathematics related to
data transmission
Most of the theory arises from earlier
developments of mathematics which were far
removed from any concrete application. One of the
main tools is the theory of finite fields, which
was invented by Galois in the XIXth century, for
solving polynomial equations by means of
radicals. The first error-correcting code
happened to occur in a sport newspaper in Finland
in 1930. The mathematical theory of information
was created half a century ago by Claude Shannon.
The mathematics behind these technical devices
are being developped in a number of scientific
centers all around the world, including in India
and in France.
http//www.math.jussieu.fr/miw/
4
French Science Today
Mathematical aspects of Coding Theory in France
The main teams in the domain are gathered in the
group C2 ''Coding Theory and Cryptography''
, which belongs to a more general group (GDR)
''Mathematical Informatics''.
http//www.math.jussieu.fr/miw/
5
French Science Today
The most important are INRIA
Rocquencourt Université de Bordeaux ENST Télécom
Bretagne Université de Limoges Université de
Marseille Université de Toulon Université de
Toulouse
http//www.math.jussieu.fr/miw/
6
INRIA
Brest
Limoges
Bordeaux
Marseille
Toulon
Toulouse
7
Institut National de Recherche en Informatique et
en Automatique
http//www-rocq.inria.fr/codes/
National Research Institute in Computer Science
and Automatic
8
Institut de Mathématiques de Bordeaux
http//www.math.u-bordeaux1.fr/maths/
Lattices and combinatorics
9
École Nationale Supérieure des Télécommunications
de Bretagne
http//departements.enst-bretagne.fr/sc/recherche/
turbo/
Turbocodes
10
Research Laboratory of LIMOGES
http//www.xlim.fr/
11
Marseille Institut de Mathématiques de Luminy
Arithmetic and Information Theory Algebraic
geometry over finite fields
12
http//grim.univ-tln.fr/
Université du Sud Toulon-Var
Boolean functions
13
Université de Toulouse Le Mirail
Algebraic geometry over finite fields
http//www.univ-tlse2.fr/grimm/algo
14
GDR IMGroupe de Recherche Informatique
Mathématique
http//www.gdr-im.fr/

• The GDR ''Mathematical Informatics'' gathers all
  the french teams which work on computer science
  problems with mathematical methods.

15
Some instances of scientific domains of the GDR
IM
http//www.gdr-im.fr/

• Calcul Formel (Symbolic computation)
• ARITH Arithmétique (Arithmetics)
• COMBALG Combinatoire algébrique (Algebraic
  Combinatorics)

16
French Science Today
Mathematical Aspects of Coding Theory in
India Indian Institute of Technology
Bombay Indian Institute of Science
Bangalore Indian Institute of Technology
Kanpur Panjab University Chandigarh University
of Delhi Delhi
17
Chandigarh
Delhi
Kanpur
Bombay
Bangalore
18
IIT BombayIndian Institute of Technology
http//www.iitb.ac.in/

• Department of Mathematics
• Department of Electrical Engineering

19
http//www.iisc.ernet.in/

• Department of Mathematics

• Finite fields and Coding Theory classification of
  permutation polynomials, study of PAPR of
  families of codes, construction of codes with low
  PAPR.

peak-to-average power
20
IIT KanpurIndian Institute of Technology
http//www.iitk.ac.in/
21
Department of Mathematics
http//www.puchd.ac.in/
22
http//www.du.ac.in/

• Department of Mathematics

23
Error Correcting Codesby Priti Shankar
http//www.ias.ac.in/resonance/

• How Numbers Protect Themselves
• The Hamming Codes Volume 2 Number 1
• Reed Solomon Codes Volume 2 Number 3

24
Playing cards
25
I know which card you selected

• Among a collection of playing cards, you select
  one without telling me which one it is.
• I ask you some questions where you answer yes or
  no.
• Then I am able to tell you which card you
  selected.

26
2 cards

• You select one of these two cards
• I ask you one question and you answer yes or no.
• I am able to tell you which card you selected.

27
2 cards one question suffices

• Question is-it this one?

28
4 cards
29
First question is-it one of these two?
30
Second question is-it one of these two ?
31
4 cards 2 questions suffice
Y Y
Y N
N Y
N N
32
8 Cards
33
First question is-it one of these?
34
Second question is-it one of these?
35
Third question is-it one of these?
36
8 Cards 3 questions
YYY
YYN
YNY
YNN
NYY
NYN
NNY
NNN
37
Yes / No

• 0 / 1
• Yin / Yang - -
• True / False
• White / Black
• / -
• Heads / Tails (tossing or flipping a coin)

38
3 questions, 8 solutions
39
Exponential law
Add one question multiply the number of cards
by 2
40
16 Cards 4 questions

• If you select one card among a set of 16, I
  shall know which one it is, once you answer my 4
  questions by yes or no.

41
(No Transcript)
42
Label the 16 cards
43
Binary representation
44
Ask the questions so that the answers are
45
The 4 questions

• Is the first digit 0 ?
• Is the second digit 0 ?
• Is the third digit 0 ?
• Is the fourth digit 0 ?

46
More difficult

• One answer may be wrong!

47
One answer may be wrong

• Consider the same problem, but you are allowed to
  give (at most) one wrong answer.
• How many questions are required so that I am able
  to know whether your answers are right or not?
  And if they are right, to know the card you
  selected?

48
Detecting one mistake

• If I ask one more question, I shall be able to
  detect if there is one of your answers which is
  not compatible with the others.
• And if you made no mistake, I shall tell you
  which is the card you selected.

49
Detecting one mistake with 2 cards

• With two cards I just repeat twice the same
  question.
• If both your answers are the same, you did not
  lie and I know which card you selected
• If your answers are not the same, I know that one
  is right and one is wrong (but I dont know which
  one is correct!).

50
4 cards
51
First question is-it one of these two?
52
Second question is-it one of these two?
53
Third question is-it one of these two?
54
4 cards 3 questions
Y Y Y
Y N N
N Y N
N N Y
55
4 cards 3 questions
0 0 0
0 1 1
1 0 1
1 1 0
56
Correct triple of answers
Wrong triple of answers
One change in a correct triple of answers yields
a wrong triple of answers
57
Boolean addition

• even even even
• even odd odd
• odd even odd
• odd odd even

• 0 0 0
• 0 1 1
• 1 0 1
• 1 1 0

58
Parity bit

• Use one more bit which is the Boolean sum of the
  previous ones.
• Now for a correct answer the sum of the bits
  should be 0.
• If there is exactly one error, the parity bit
  will detect it the sum of the bits will be 1
  instead of 0.

59
8 Cards
60
4 questions for 8 cards
Use the 3 previous questions plus the parity bit
question
61
First question is-it one of these?
62
Second question is-it one of these?
63
Third question is-it one of these?
64
Fourth question is-it one of these?
65
16 cards, at most one wrong answer 5 questions
to detect the mistake
66
Ask the 5 questions so that the answers are
67
Correcting one mistake

• Again I ask you questions where your answer is
  yes or no, again you are allowed to give at most
  one wrong answer, but now I want to be able to
  know which card you selected - and also to tell
  you whether and when you lied.

68
With 2 cards

• I repeat the same question three times.
• The most frequent answer is the right one vote
  with the majority.
• 2 cards, 3 questions, corrects 1 error.

69
With 4 cards
If I repeat my two questions three times each, I
need 6 questions
Better way repeat each of the two questions
twice only, and use the parity check bit. 5
questions suffice
70
4 right answers

• 0 0 0 0 0 or Y Y Y Y Y
• 0 1 0 1 1 or Y N Y N N
• 1 0 1 0 1 or N Y N Y N
• 1 1 1 1 0 or N N N N Y
• If there is just one mistake, it is easy to
  correct it.
• 4 cards, 5 questions, corrects 1 error.

71
With 8 Cards
If I repeat 3 times my 3 questions, I need 9
questions
With 6 questions only I can correct one error
72
8 cards, 6 questions, corrects 1 error

• Ask the three questions giving the right answer
  if there is no error, then use the parity check
  for questions (1,2), (1,3) and (2,3).
• Right answers
• (a, b, c, ab, ac, bc)
• with a, b, c replaced by 0 or 1

73
Number of questions
74
With 16 cards, 7 questions suffice to correct
one mistake
75
Error correcting codes
76
Coding Theory

• Coding theory is the branch of mathematics
  concerned with transmitting data across noisy
  channels and recovering the message. Coding
  theory is about making messages easy to read
  don't confuse it with cryptography which is the
  art of making messages hard to read!

77
Claude Shannon

• In 1948, Claude Shannon, working at Bell
  Laboratories in the USA, inaugurated the whole
  subject of coding theory by showing that it was
  possible to encode messages in such a way that
  the number of extra bits transmitted was as small
  as possible. Unfortunately his proof did not give
  any explicit recipes for these optimal codes.

78
Richard Hamming

• Around the same time, Richard Hamming, also at
  Bell Labs, was using machines with lamps and
  relays having an error detecting code. The digits
  from 1 to 9 were send on ramps of 5 lamps with
  two on and three out. There were very frequent
  errors which were easy to detect and then one had
  to restart the process.

79
The first correcting codes

• For his researches, Hamming was allowed to have
  the machine working during the week-end only, and
  they were on the automatic mode. At each error
  the machine stopped until the next monday
  morning.
• "If it can detect the error," complained
  Hamming, "why can't it correct it! "

80
The origin of Hammings code

• He decided to find a device so that the machine
  would not only detect the errors but also correct
  them.
• In 1950, he published details of his work on
  explicit error-correcting codes with information
  transmission rates more efficient than simple
  repetition.
• His first attempt produced a code in which four
  data bits were followed by three check bits which
  allowed not only the detection, but also the
  correction of a single error.

81
(No Transcript)
82
Codes and Geometry

• 1949 Marcel Golay (specialist of radars)
  produced two remarkably efficient codes.
• Eruptions on Io (Jupiters volcanic moon)
• 1963 John Leech uses Golays ideas for sphere
  packing in dimension 24 - classification of
  finite simple groups
• 1971 no other perfect code than the two found by
  Golay.

83
Error Correcting Codes Data Transmission

• Telephone
• CD or DVD
• Image transmission
• Sending information through the Internet
• Radio control of satellites

84
Applications of error correcting codes

• Transmitions by satellites
• Compact discs
• Cellular phones

85

• Olympus Mons on Mars Planet
• Image from Mariner 2 in 1971.

86

• Between 1969 and 1973 the NASA Mariner probes
  used a powerful Reed-Muller code capable of
  correcting 7 errors out of 32 bits transmitted,
  consisting now of 6 data bits and 26 check bits!
  Over 16,000 bits per second were relayed back to
  Earth.

The North polar cap of Mars, taken by Mariner 9
in 1972.
87
Voyager 1 and 2 (1977)

• Journey Cape Canaveral, Jupiter, Saturn, Uranus,
  Neptune.
• Sent information by means of a binary code which
  corrected 3 errors on words of length 24.

88
Mariner spacecraft 9 (1979)

• Sent black and white photographs of Mars
• Grid of 600 by 600, each pixel being assigned one
  of 64 brightness levels
• Reed-Muller code with 64 words of 32 letters,
  minimal distance 16, correcting 7 errors, rate
  3/16

89
Voyager (1979-81)

• Color photos of Jupiter and Saturn
• Golay code with 4096212 words of 24 letters,
  minimal distance 8, corrects 3 errors, rate 1/2.
• 1998 lost of control of Soho satellite recovered
  thanks to double correction by turbo code.

90
NASA's Pathfinder mission on Mars

• The power of the radio transmitters on these
  craft is only a few watts, yet this information
  is reliably transmitted across hundreds of
  millions of miles without being completely
  swamped by noise.

Sojourner rover and Mars Pathfinder lander
91
Listening to a CD

• On a CD as well as on a computer, each sound is
  coded by a sequence of 0s and 1s, grouped in
  octets
• Further octets are added which detect and correct
  small mistakes.
• In a CD, two codes join forces and manage to
  handle situations with vast number of errors.

92
Coding the sound on a CD
On CDs the signal in encoded digitally. To guard
against scratches, cracks and similar damage,
two "interleaved" codes which can correct up to
4,000 consecutive errors (about 2.5 mm of track)
are used.

• Using a finite field with 256 elements, it is
  possible to correct 2 errors in each word of 32
  octets with 4 control octets for 28 information
  octets.

93
A CD of high quality may have more than 500
000 errors!

• After processing of the signal in the CD player,
  these errors do not lead to any disturbing noise.
• Without error-correcting codes, there would be no
  CD.

94
1 second of audio signal 1 411 200 bits

• 1980s, agreement between Sony and Philips norm
  for storage of data on audio CDs.
• 44 100 times per second, 16 bits in each of the
  two stereo channels

95
Codes and Mathematics

• Algebra
• (discrete mathematics finite fields, linear
  algebra,)
• Geometry
• Probability and statistics

96
Finite fields and coding theory

• Solving algebraic equations with
  radicals Finite fields theory
  Evariste Galois
  (1811-1832)
• Construction of regular polygons with rule and
  compass
• Group theory

Srinivasa Ramanujan (1887-1920)
97
Coding Theory
98
Error correcting codes
99
Principle of coding theory

• Only certain words are allowed (code
  dictionary of valid words).
• The  useful  letters (data bits) carry the
  information, the other ones (control or check
  bits) allow detecting or correcting errors.

100
Detecting one error by sending twice the message

• Send twice each bit
• 2 code words among 422 possible words
• (1 data bit, 1 check bit)

• Code words
• (two letters)
• 0 0
• and
• 1 1
• Rate 1/2

101

• Principle of codes detecting one error
• Two distinct code words
• have at least two distinct letters

102
Detecting one error with the parity bit

• Code words (three letters)
• 0 0 0
• 0 1 1
• 1 0 1
• 1 1 0
• Parity bit (x y z) with zxy.
• 4 code words (among 8 words with 3 letters),
• 2 data bits, 1 check bit.
• Rate 2/3

2
103
Code Words Non Code Words

• 0 0 0 0 0 1
• 0 1 1 0 1 0
• 1 0 1 1 0 0
• 1 1 0 1 1 1
• Two distinct code words
• have at least two distinct letters.

2
104
Check bit

• In the International Standard Book Number (ISBN)
  system used to identify books, the last of the
  ten-digit number is a check bit.
• The Chemical Abstracts Service (CAS) method of
  identifying chemical compounds, the United States
  Postal Service (USPS) use check digits.
• Modems, computer memory chips compute checksums.
• One or more check digits are commonly embedded in
  credit card numbers.

105
Correcting one errorby repeating three times

• Code words
• (three letters)
• 0 0 0
• 1 1 1
• Rate 1/3

• Send each bit three times
• 2 code words
• among 8 possible ones
• (1 data bit, 2 check bits)

106

• Correct 0 0 1 as 0 0 0
• 0 1 0 as 0 0 0
• 1 0 0 as 0 0 0
• and
• 1 1 0 as 1 1 1
• 1 0 1 as 1 1 1
• 0 1 1 as 1 1 1

107

• Principle of codes correcting one error
• Two distinct code words have at least three
  distinct letters

108
Hamming Distance between two words

• number of places in which the two words
  
• differ
• Examples
• (0,0,1) and (0,0,0) have distance 1
• (1,0,1) and (1,1,0) have distance 2
• (0,0,1) and (1,1,0) have distance 3
• Richard W. Hamming (1915-1998)

109
Hamming distance 1
110
Hammings unit sphere

• The unit sphere around a word includes the words
  at distance at most 1

111
At most one error
112
Words at distance at least 3
113
Decoding
114
The code (0 0 0) (1 1 1)

• The set of words with three letters (eight
  elements) splits into two balls
• The centers are (0,0,0) and (1,1,1)
• Each of the two balls consists of elements at
  distance at most 1 from the center

115
Two or three 0
Two or three 1
(0,0,1)
(1,0,1)
(0,1,0)
(1,1,0)
(0,0,0)
(1,1,1)
(1,0,0)
(0,1,1)
116
2 data bits, 3 check bits, corrects 1 error

• Code words a b a b ab
• 0 0 0 0 0
• 0 1 0 1 1
• 1 0 1 0 1
• 1 1 1 1 0
• Two code words have distance at least 3
• Rate 2/5.

117
3 data bits, 3 check bits, corrects 1 error

• Code words a b c ab ac bc
• 0 0 0 0 0 0 1 0 0 1 1 0
• 0 0 1 0 1 1 1 0 1 1 0 1
• 0 1 0 1 0 1 1 1 0 0 1 1
• 0 1 1 1 1 0 1 1 1 0 0 0
• Two code words have distance at least 3
• Rate 1/2.

118
4 data bits, 3 check bits, corrects 1 error

• Hammings code, 1950
• Rate 4/7.

Generalization of the parity check bit
119
Hammings code
120
How to compute e , f , g from a , b , c , d.
eabd
d
a
b
facd
c
gabc
121
Hamming code

• Words of 7 letters
• Code words (1624 among 12827)
• (a b c d e f g)
• with
• eabd
• facd
• gabc
• Rate 4/7

122
16 code words of 7 letters

• 0 0 0 0 0 0 0
• 0 0 0 1 1 1 0
• 0 0 1 0 0 1 1
• 0 0 1 1 1 0 1
• 0 1 0 0 1 0 1
• 0 1 0 1 0 1 1
• 0 1 1 0 1 1 0
• 0 1 1 1 0 0 0

• 1 0 0 0 1 1 1
• 1 0 0 1 0 0 1
• 1 0 1 0 1 0 0
• 1 0 1 1 0 1 0
• 1 1 0 0 0 1 0
• 1 1 0 1 1 0 0
• 1 1 1 0 0 0 1
• 1 1 1 1 1 1 1

Two distinct code words have at least three
distinct letters
123
The binary code of Hamming (1950)

• It is a linear code (the sum of two code words
  is a code word) and the 16 balls of radius 1 with
  centers in the code words cover all the space of
  the 128 binary words of length 7
• (each word has 7 neighbors (71)?16 256).

124
Playing cards
125
7 questions to find the selected card among 16,
with one possible wrong answer

• Replace the cards by labels from 0 to 15 and
  write the binary expansions of these
• 0000, 0001, 0010, 0011
• 0100, 0101, 0110, 0111
• 1000, 1001, 1010, 1011
• 1100, 1101, 1110, 1111
• Using the Hamming code, get 7 digits.
• Select the questions so that Yes0 and No1

126
7 questions to find the selected number in
0,1,2,,15 with one possible wrong answer

• Is the first binary digit 1?
• Is the second binary digit 1?
• Is the third binary digit 1?
• Is the fourth binary digit 1?
• Is the number in 1,2,4,7,9,10,12,15?
• Is the number in 1,2,5,6,8,11,12,15?
• Is the number in 1,3,4,6,8,10,13,15?

127
The Hat Problem
128
The Hat Problem

• Three people are in a room, each has a hat on his
  head, the color of which is black or white. Hat
  colors are chosen randomly. Everybody sees the
  color of the hat on everyones head, but not on
  their own. People do not communicate with each
  other.
• Everyone gets to guess (by writing on a piece of
  paper) the color of their hat. They may write
  Black/White/Abstain.

129
Rules of the game

• The people in the room win together or lose
  together.
• The team wins if at least one of the three people
  did not abstain, and everyone who did not abstain
  guessed the color of their hat correctly.
• How will this team decide a good strategy with a
  high probability of winning?

130
Strategy

• Simple strategy they agree that two of them
  abstain and the other guesses randomly.
• Probability of winning 1/2.
• Is it possible to do better?

131
Information is the key

• Hint
• Improve the odds by using the available
  information everybody sees the color of the hat
  on everyones head but himself.

132
Solution of the Hat Problem

• Better strategy if a member sees two different
  colors, he abstains. If he sees the same color
  twice, he guesses that his hat has the other
  color.

133

• The two people with white hats see one white
  hat and one black hat, so they abstain.

The one with a black hat sees two white hats,
so he writes black.
They win!
134

• The two people with black hats see one white
  hat and one black hat, so they abstain.

The one with a white hat sees two black hats,
so he writes white.
They win!
135
Everybody sees two white hats, and therefore
writes black on the paper.

• They lose

136
Everybody sees two black hats, and therefore
writes white on the paper.

• They lose

137

• Winning

two white or two black
138

• Losing

three white or three black
139

• The team wins exactly when the three hats do not
  have all the same color, that is in 6 cases out
  of a total of 8
• Probability of winning 3/4.

140
Connection with error detecting codes

• Replace white by 0 and black by 1
• hence the distribution of colors becomes a
  word of three letters on the alphabet 0 , 1
• Consider the centers of the balls (0,0,0) and
  (1,1,1).
• The team bets that the distribution of colors is
  not one of the two centers.

141
Assume the distribution of hats does not
correspond to one of the centers (0, 0, 0) and
(1, 1, 1). Then

• One color occurs exactly twice (the word has both
  digits 0 and 1).
• Exactly one member of the team sees twice the
  same color this corresponds to 0 0 in case he
  sees two white hats, 1 1 in case he sees two
  black hats.
• Hence he knows the center of the ball (0 , 0 ,
  0) in the first case, (1, 1, 1) in the second
  case.
• He bets the missing digit does not yield the
  center.

142

• The two others see two different colors, hence
  they do not know the center of the ball. They
  abstain.
• Therefore the team wins when the distribution of
  colors does not correspond to one of the centers
  of the balls.
• This is why the team wins in 6 cases.

143

• Now if the word corresponding to the distribution
  of the hats is one of the centers, all members of
  the team bet the wrong answer!
• They lose in 2 cases.

144
Hat problem with 7 people
For 7 people in the room in place of 3, which is
the best strategy and its probability of
winning?
Answer the best strategy gives a probability
of winning of 7/8
145
The Hat Problem with 7 people

• The team bets that the distribution of the hats
  does not correspond to the 16 elements of the
  Hamming code
• Loses in 16 cases (they all fail)
• Wins in 128-16112 cases (one bets correctly, the
  6 others abstain)
• Probability of winning 112/1287/8

146
SPORT TOTO the oldest error correcting code

• A match between two players (or teams) may give
  three possible results either player 1 wins, or
  player 2 wins, or else there is a draw (write 0).
• There is a lottery, and a winning ticket needs to
  have at least three correct bets. How many
  tickets should one buy to be sure to win?

147
4 matches, 3 correct forecasts

• For 4 matches, there are 34 81 possibilities.
• A bet on 4 matches is a sequence of 4 symbols
  0, 1, 2. Each such ticket has exactly 3 correct
  answers 8 times.
• Hence each ticket is winning in 9 cases.
• Since 9 ? 9 81, a minimum of 9 tickets is
  required to be sure to win.

148
9 tickets

• 0 0 0 0 1 0 1 2 2 0 2 1
• 0 1 1 1 1 1 2 0 2 1 0 2
• 0 2 2 2 1 2 0 1 2 2 1 0

Rule a b ab a2b modulo 3
This is an error correcting code on the
alphabet 0, 1, 2 with rate 1/2
149
Sphere Packing

• While Shannon and Hamming were working on
  information transmission in the States, John
  Leech invented similar codes while working on
  Group Theory at Cambridge. This research included
  work on the sphere packing problem and culminated
  in the remarkable, 24-dimensional Leech lattice,
  the study of which was a key element in the
  programme to understand and classify finite
  symmetry groups.

150
Sphere packing
The kissing number is 12
151
Sphere Packing

• Kepler Problem maximal density of a packing of
  identical sphères
•   p / Ö 18 0.740 480 49
• Conjectured in 1611.
• Proved in 1999 by Thomas Hales.
• Connections with crystallography.

152
Current trends

• In the past two years the goal of finding
  explicit codes which reach the limits predicted
  by Shannon's original work has been achieved. The
  constructions require techniques from a
  surprisingly wide range of pure mathematics
  linear algebra, the theory of fields and
  algebraic geometry all play a vital role. Not
  only has coding theory helped to solve problems
  of vital importance in the world outside
  mathematics, it has enriched other branches of
  mathematics, with new problems as well as new
  solutions.

153
Directions of research

• Theoretical questions of existence of specific
  codes
• connection with cryptography
• lattices and combinatoric designs
• algebraic geometry over finite fields
• equations over finite fields

154
Some recent results in mathematics related to
data transmission

• Michel Waldschmidt
• Université P. et M. Curie - Paris VI
• Centre International de Mathématiques Pures et
  Appliquées - CIMPA

India, October-November 2007
http//www.math.jussieu.fr/miw/