Social Processes and Proofs of Theorems and Programs reaction paper Read Social Processes and Proofs of Theorems and Programs by DeMillo, Lipton and Perlis

Social Processes and Proofs of Theorems and Programs reaction paper Read Social Processes and Proofs of Theorems and Programs by DeMillo, Lipton and Perlis. Write a 600 – 700 word reaction paper and discuss the following: mathematics and proof as a formal process, mathematics and proof as a social process, the classical view versus the probabilistic view of mathematical belief, lessons learned from mathematical proof and believability that can be applied to the problem of software correctness soclAL
pROCESSES
AND
pROOFS
Richard
Institute
Georgia
Richard
J.
Lipton
Yale
“I
should
that
like
to
Descartes
to
give
ask
a precise
which
as
intuitive
my vague
logical
correctness.
to
that
r,
.
.
.
not
is
final
the
is
be
are
average
the
and
that
for
you
authority
.
.
.
J.
and
science
more
like
that
this
.“
Barkley
Rosser
It
is
exactly
of
that
program
large
is
of
to
computer
in
its
software.
shocking
rests
on
large
measure
of
of
mechanism
of
Many
of
ference
in our
This
credit
on
and
work.
has
work
was
key
to
the
our
in
discussions
social
this
paper
at
in
were
the
12
Office,
29–76–G–0338
Grant
and
by
in
No.
NSF
by
Authors’
of
Addresses:
and Computer
Technology,
Perlis
versity,
‘Logic
,
‘proof’
sense
of
proof
that
Grant
No.
R. DeMillo,
Science,
Atlanta,
Department
New Haven,
for
GA 30332;
may
DAAG
of Computer
CT 06520.
Mathematicians,
School
of InGeorgia
Institute
Lipton,
Science,
Yale
be
that
an
destructive
part
as
of
this
positive
software
we
that
make
and
conclude
believe
can
that
there
program
are
verifica-
mathematics.
‘proving’
in
to
The
the
everyday
two
different
doubly
common
to
in
the
notion
the
senses
quoted
mathematics;
refers
appears
a proo~we
A.
sense
of
the
formal
informal
italicized
notion
literature
carried
1953.
or
Peano
to
confuse
call
this
is
easily
206
out
tioned
view
mean
syntactically
theory
matics
Uni–
shall
calculus
hypotheses)
McGraw–Hill,
to
paragraph.
in
formalistic
second
and
that
“real”
step-by–step,
DcR74–12870.
R.
reason
used
the
reliable
the
view
process,
arguments
the
In
and
the
tools,
‘proof’
By
US Army
DAHC04–74–G-0179,
these
refers
used
& 13,
particular
this
con–
the
In
that
formal
misleading
become
argue
on
of
program
verification.
3?
formation
of
proof
DOD
encouragement
part
is
producing
like
a previous
logical
supported
that
every
in
our
and
mechanisms
more
present
process
for
We have
proving
July
continuin~
tion
success
the
NC.
We acknowledge
Suttle
who organized
of
to
wrong.
a
interpret
is
position
held
Technology,
been
there
social
we
but
are
than
programming
the
A
to
in
better
found
continued
belongs
reported
of
have
paper
we
correct,
it
art
to
paper
rather
software.
we
of
observation.
for
Software
Durham,
of J.R.
we
theorems;
ideas
out
Research
the
simple
“proving”
the
Conference
1976
in
the
help
but
that
mathematics
crystallized
in
piece
the
is
the
strive
this
for
mathematics
suggestions
to
confidence
a view
many,
purpose:
view
given
social
paper
a
are
that
should
In
of
of
proving
argued
programming
is,
of
for
mechanisms
extensively
point
part
view
community,
a particular
relatively
growth
+
is
to
a
primary
of
This
mediate
by
of
mathematics
require
science
one’s
functioning
which
that
as perceived
increase
correct
processes
mathematics
the
fail
dramatically
and
in
verification,
segment
bound
be
those
theorems
social
mathematics.
ongoing
proofs
been
usually
first
+
these
lacking.
has
reasons
intuition
.
however,
It
intend
.
Perlis*
totally
same
same?
his
J.
almost
mathematician]
forget
Alan
programs,
logical
the
+
PROGRAMS
University
proposing
of
to
AND
DeMillo*
of Technology
question
are
feeling
How do
they
[the
should
same
You
definition
correctness
show
the
asked.
A.
OF THEOREMS
one
within
proceeds
to
justifiable
as
proofs
a
We have
axioms
from
by
its
set
been
condi–
Let
The
monolithic
school
of
(or
postulates
steps,
each
predecessors
as
formal,
“proofs.”
monolithic.
theorems
valid,
deduction
consistent,
with
Hilbert
from
of
Zermelo-Fraenkel
arithmetic.
of
sort
checkable
such
viewpoint
the
the
logic:
us
mathe–
or
of
which
by
a
is
fixed
allowable
by
rule
arguing
can
be
(and
of
that
very
transformation.
the
views
misleading
must
be)
and
highly
every
step
of
which
to
gain
that
the
it
is
have
particular
led
of
and
into
the
has
not
the
other
over
Ulam
200,000
cians.
theorems
A fair
but
or
are
a
of
work
of
that
discredited
tent
a
the
are
not
“proof”
stood
the
of
for
was
fallacious
managed
by
to
at
problem
only
be
contains
a
detected
by
induction
involving
Mathematical
famous
mathematicians
sometimes
less
Legendre
gress
rising
only
pocket,
to’
and
so
a very
to
pause,
quietly
they
problem).
an
subtle
lemmas
chained
filled
with
is
best
perfect––of,
address
return
announce
of
that
the
i+
is
+
HThe
for
this
Architecture
57 (1950):
authors
are
of Mathematics,”
221-323.
indebted
to
get
of
The
the
the
Notice
issue.
theorem,
tend
weight
“proof,”
compared
to
to
at
the
this
far
regard
that
more
as
mathematicians
for,
of
theo-
was
he
has
presumably,
a
so
since
career
labor
can
be
Science
[5]
contains
that
believed,
be
the
in
judged
more
may
issue
here
are
secure
due
is
for
not
what
his
There
roughtly
be
rather
asserts
a
apparently
truth
but
one
theorem.
The
con-
Y.
the
central
when
may
his
how
it
belief
two
classified
CZassieal
mathematical
Math.
there
a suitable
completely
Zalcsteim
is,
observation.
207
and
View.
statement
principle,
would
truth:
Prof.
reconsider
his
supporting
Americans.
we
issue
The
believed
that
instance,
American
the
researchers
time
is
of
that
a
relevant
views
classical
as
and
probabilistic.
in
The
Monthly,
of
mathematical
particular
of
matter
“proof,”
this
suggestion
of
“proofs”
an
were
to
‘
others
group
low
of
moti-
third
experience
results
ex–
highly
“proof,”
one
American
this
the
the
their
mathematics
revision.
together.
efforts
notes
the
notion
can
a mathematical
his
up
provocative
stories
for
given
the
A recent
important
analysis
by
which
300-
it
of
in
against
very
were
was
incor–
results
One
“proofs”
issue,
clearly
truth
still
the
objectively.
The
the
that
careful
whose
than
who
a monumental
is
press
the
was
of
to
it
But
“proof”
being
probable
the
unsettled
been
that
favor
now
comprehensible
(this
[4]–unfortunately,
flaw
that
rems.
mathematicians
settling
many
folklore
[3]
in
to
nor
a
retired
point,
in
since
result
calculation,
group
yet
another
evidence
Japanese
published
flaw
has
celebrated
theory
are
which
colleagues
published
group
[2]
Heawood
way,
the
of
yet
since
American
discredited;
Subsequently,
and
involved
goofed.
American
error
in
The
and
high
Japanese
the
arising
topology).
numerical
an
the
indepenthe
contradictory,
each
discover
be
rect.
incompe–
fatal
well-known
apparently
i.n
“proof”
the
this
Britton
“proof,”
a
by
their
solved
algebraist
page
many
convince
last
Kempe
before
by
of
to
neither
two
concerning
complex
sufficiently
and
necessarily
object
who.had
Obviously
obtained
the
that
conjecture
uncovered
“proof,”
rediscovered
had
years
those
Japanese
the
could
a
seldom
work
1879,
four-color
eleven
reasoning
the
In
the
are
so
not
American,
of
of
be
evident
were
vated
community.
and
always
mathematicians.
by
to
and
all
stakes
but
ignored,
believed
published
scientists,
mathemati–
are
at
type
very
symbolic
not
changed.
reasons,
mathematical
get
crackpot
other
be
year
subsequently
most
to
every
by
are
doubt,
come
portion
Theorems
these
for
into
fraction
sizea,ble
of
-t
out
were
mathematicians
results
(a
a branch
turned
detailed
Bourbaki
that
theory;
“proofs”
was
published
disallowed
thrown
tiny
estimates
are
number
contradicted
others
[1]
in
one.”
(one
announced
group
of
Recently
topologists
Japanese)
results
does
“proofs.”
of
homotopy
both
of
deductions
N.
Stanislaw
tried
number
problem
believable
homotopy
ideas
the
a given
groups
same
deductions
construction
of
every
on
other
that
if
one
step
the
insight
to
to
knows
thought.
increasing
working
dent
‘understood?
than
verify
chain
preference
is
insure
composed
a clear
this
Just
THEOREMS
been
more
further
school
mathematics
correctness
which
require
begin
Hilbert
mathematician
a Proof
has not
has
done
nothing
by
the
non-monolithic.
BELIEVING
“Indeed,
We will
of
is
When
A,
one
a correct,
logical
formalizes
the
is
true
if
it
says
of
which
believes
that
says
is
not
in
pPOOf
of
A
(semantically)
Aristotelian
it
a
that,
formal
theory
‘“A’
one
believes
of
notion
what
that
is
it
of
that
is
not.”
it
Two
points
the
classical
ties
of
lent
to
deserve
view
fashion:
objects
counterpart
In
.
even
produce
such
for
our
mathematical
theories,
proofs
finite
outstanding
such
research
physical
prove
as
[6]
binary
strings
will
of
even
allowing
logy
that
observable
with
an
of
a computer
in
set
sketch
of
infinitely
thin
densely
fill
precise
.
proofs
coin–
amount
of
This
we
.
.
let
us assume
. .“
.
without
loss
of
replace
therefore,
. . .“
enormous
by
classical
deliberately
of
mathematical
that
the
can
is
dual
a
isolate
classical
only
of
tion
such
as
in
leads
of
critical
classical
to
a more
“proofs,”
viewed
as
reasonable
The
reasoning
“proof”
technically
may
allow
ideas
into
give
in
cians
rise
The
to
are
(classically)
208
alternative
the
same
of
others’
offices
They
key
is
to
theorem.
An
a more
symposium
that
and
doubting
arguments,
with
they
news
of
scribble
on
expensive
convincing
other
other
mathemati-
listen!
“proof”
when
transformation
the’
publica–
mathemati-
and
.
of
after
convince
and
inclined
“proofs”
must
Finally,
their
of
con–
audience
give
to
sake
iamediate
to
They
the
most
be
takes
audiences
a
and
reviewed.
Third,
cafeterias
to
is
mechanism.
must
Reviews
for
happens
it
a reviewing
research,
All
What
and
attempt
university
for
students,
article.
other.
it
a second
“proof.”
each
internalization
this
to
cleared,
current
.
us
mathema–
“proof,”
their
A large
which
mathematicians
have
mathematicians
that
for
restaurants
easier
“proofs’’-–hopefully,
plausible
hostile)
burst
napkins
is
may
(sometimes
they
be
author
reads
A referee
the
each
talks
insights
to
demonstrations
being
and
detail.
to
colloquium
its
well
several
published
has
at
to
An
and
same
leads
Mat7zematieaZ
look
in
Let
read.
by
a printed
smoke
talk
the
all,
publication.
the
this
certainly
theorems
that
filter.
in
that
exceedingly
paper
case
refereed,
before
some
them
triple
and
Almost
widely
on
This
convinced
cians
long
be
perhaps
valid
one
those
uncertainty
process
Since
statements.
advantag@
if
formal
probabilistic
a probabilistically
than
to
it
give
of
Vieti.
view,
correct,
renumbering
amounts
T?w ProbzbiZistie
probably
a
leisurely
.
are
on
by
are
is
vinced
be
!,
. .
necessary
the
others.
many
of
and
in
by
the
generality
to
collaborate
read
impor-
mathematicians.
common
ticians
that
success–
mathematics
cause
of
time.
mechanisms.
down
the
usually
about
example,
In
that
survive?
operate
these
“proof”
.
lengths
For
of
“proofs”
a
errors
possible
that
communities
writes
it
and
such
it
so
Obviously,
understood
some
of
is
mathematicians
not
mechanisms
First,
techno–
end,
any
amount
been
is
insight
probabili-
by
Is
than
that
small
nature?
believed
will
and
large
Then
electronic
These
The
believed
a
form
not.
stateformulas
bits.
at
classical
test
Rabin’s
has
of
computational
how
and
of
is
must
“proofs.”
a description
process
to
guarantee
survived
survive.
be
to
all
whether
the
difficulties,
these
may
for
earth.
a reasonable
of
intuitions
in
has
N,
of
vanishingly
these
mOst
use
!!
as
concerning
our
of
mathematics
ful
with
within
view
integers
the
is
probable
because
possible
not
error,
“Proofs”
embedded
often
of
In
trivial
is
or
of
1000
universe.
it
prime
notion
determining
required
of
history
build
with
that
is
lifetime
get
a simple
design
the
theorems
it
luxury
them
we
ty
time
in
N is
large
fail
this
testing
for
composite
tant
proton–sized
connecting
was
to
a striking
most
the
does
very
Albert
hard
to
of
left
statements
logical
out
truth
produce
observations
detail
set
For
for
techniques
N is
of
algorithm
[7].
not
illustration
long.
mathematical
say,at
computer
entire
with
encode
ourselves
will
the
tide
the
length,
components,
the
we
and
decide
formulas
wires,
that
the
how
simple
Suppose
simple
Rabin’s
reasons–more
every
most
An
classical
formalize
the
on
of
comparatively
.
that
are
to

impossibly
concludes
or
possible
gods
there
lecture
fails
the
a mathematically
even
interpretation
ments
is
For
are
primality
formal
view
humanly
the
arguments.
its
classical
There
biva–
or
that
by
be
allowing
whose
Meyer’s
it
Michael
Secondly,
require
the
that
a proof.
reason
of
not
“proof.”
proper–
strictly
object.
accompanied
fact,
require
a
judge
ascribes
an
does
be
in
valid
First,
here.
we
either
to
view
“proof”
some
that
a theorem
classical
sound
requires
a property
informal
not
attention
mathematical
ascribe
the
special
is
“proof.”
and
excellent
it
ThJs
several
source
is
believed?
probably
leads
versions
of
an
usually
of
examples
of
the
this
transformations
process
and
Joel
ties
Spencer
of
[8].
in
a slightly
community
concepts
are
cal
an
truth
that
If,
in
of
the
as
not
lead
to
in
the
supporting
increased.
example,
in”
values
and
relying
bridges
areas
invariably
do–and
information
matics
most
prime
connections
Dirichlet
certain
prime
of
the
is
notion
Cohen’s
independence
of
Generalized
axioms
of
believed
Dana
set
(i.e.,
Scott,
the
ultimate
theorems
this
refereed,
discussed
“plugging
4.
internalized
physical
5.
generalized
6.
used,
7.
connected
in
of
Solovay
of
of
Choice
that
by
few
J.
and
seems
this
believe
formalized
do
that
theory
Rather,
because
not
they
logical
and
and
and
mathe–
they
are:
reviewed
paraphrased
other
theorems.
ROLE
OF SIMPLICITY
that
very
rather
logic.
remaining
then
Barkley
it
be
take
Every
If
through
a
its
An
are
important
Rosser
––-
“clean”
theorem
is
209
is
a –-–––,
and
–-–-
for
and
—–
and
special.cases
s
unless
i.
————-,
.
—-—
the
form
-–––-–
—-except
i.
————.
every
its
rule,
or
~iv.———–
logicians.
is
with
a general
problems
the
7 listed
theorem
importance,
correlated
As
are
form
xx—————
was
theorems
which
by
highly
state.
the
Theorem.
and
1
to
mathematical
to
than
filters
degree
statement.
to
likely
few
governed
to
of
simple
the
the
largely
!lkeorwrn.
of
same
of
important
the
the
radical
and
is
the
and
from
very
this
usually
believed
with
azz
believed
more
digestion
of
Indeed,
and
the
mathematical
to
above.
most
and
sequence
rapid
[10]
so
of
analysis
the
of
and
subject
simplicity
“proof”
Riemann
Hypothesis
was
mathe-
“techno-
[9]
example
“proof”
Axiom
of
complex
into
of
celebrated
by
argument
published,
We acknowledge
theorems.
“proof”
in
Symbolic
contact
of
sort
examined
arbiter.
get
3.
that
a
certain
forcing).
they
to
in
of
mathematicians
because
in
read
1.
ones
in
recent
understood)
Robert
of
main
arguments
students
papers
the
logicians
forcing
Journal
use
them
as
the
the
the
National
tools
summarize
2.
by
of
ideas
found
When
few
graduate
essential
“proofs”
matical
ways.
transferal
This
@Z&ng
theory
us
the
very
1976,
con–
familiar
useful.
standard
from
Working
act
is
other
as
(half
translate
confidence
make
comparatively
the
to
,
they
used
by
[12]
and
to
By
THE
properties
Continuum
make
can
They
if
branches
of
A more
of
are
more
1964,
“proofs.”
logic
believe
results
in
selection
Let
“proof”
successful
theorems
original
Logic
mathematics––important
theorem.
the
way.
impressive
confidence
asymptotic
of
random
of
his
others
concept,
immensely
studied
section.
Planes
are
established
numbers.
phenomenon
Paul
theorems
Hadamard’s
between
areas
place
used.
in
a particular
distinct
is
number
a larger
used
on
example
transfer”
of
sometimes
increases
famous
and
the
the
his
logic
be
Sciences
routinely
classi-
truth
are
of
theorems.
the
between
also
uses
of
some
“proofs”;
use
theorems
other
are
with
forc~ng
of
characteriza-
and
the
to
believed
con-
established.
be
stand
in
the
of
conclusion.
that
belief
Believable
the
the
the
Robinson
arguments
announced
notion
algebraic
generalized
generalization
the
then
can
of
of
probable
larger
lemmas
interpretation
evidence
the
theorems
engineers
the
in
contradictions,
and
If
takes
belief
forcing
Cohen
central
alternative
Abraham
logic,
Academy
considered
in
in
established.
the
is
netted
central
and
an
[11].
the
generalization,
then
Theorems
For
logy
to
lemmas
do
The
lead
important
appear
with
the
statement
Finally,
may
inter–
sometimes
generalized
to
“proof,”
original
fly
is
mechanisms
enough
Cohen’s
it
construction,
angles,
becomes
that
a
intuition
right”
enough
theorem
theorem
generalized
“feel
from
passage
that
giving
tion
in
the
stab~~ity.
to
activity
the
After
decide
Erd6s
proper–
aids
“proof”
internalized
by
Paul
presented
which
ultimate
the
next
the
social
of
way.
begin
examined
The
is
each
to
place
by
combinatorial
usually
alternative
proofs
cepts
are
seems
have
various
The
[8]
take
monograph
different
and
can
little
guises,
nalization
the
that
the
discussed
variety
in
is
that
satisfies
—
is
a —–
—-
much
The
problems
icians
,,
since
In”fact,
that
the
tail
is
often
that
the
man
by
on
the
aPPlies
be
to
consider
that
it
can
child.
location
of
may
not
certainly
tion
“The
to
of
example,
The
four-
?
Even
be
precise
from
precision
asks
tion
to
in
A simple
statement
complicated
theory;
but
this
sort
the
social
we
are
of
in
rists.
But
their
ized
be
in
again,
read,
chratic,
paltry
will
exactly
a
look
to
set
of
say,
ever
this
that
“grunt
be
“grunt
that
and
consider
kind
easy
to
that
of
the
theorem
no
idiosyn–
to
such
a
Yet
arises
it
in
the
rather
the
proofs
of
Fidelity
One
pro-
Really
proving.
For
real
programs
deal
with
real
210
even
in
know
of
proposing
by
teams
these
are
verification
sensibilities,
is
exactly
why
American
mechanisms
not
are
proving.
basic,
programs
Thus
,
the
apply
and
run
into
incentive
Discourse:
Math,
are
will
nor
simple
Mathematical
Two?”
th…
Purchase answer to see full
attachment