1 Władysław M. Tur ski
Mar 26, 2015
1 Władysław M. Turski
Władysław M. Turski 2
It Was Fun
While It Lasted
Władysław M. Turski 3
It Wasn't Much Fun
1959, Moscow (USSR), M.Sc. Thesis
I was a computer
Władysław M. Turski 4
It Was Fun
1960, Manchester (UK), Mercury
When love prevails
Logic fails
So the computer
Was made neuter
Władysław M. Turski 5
It Was Fun
Computing meteor stream orbits
Władysław M. Turski 6
It Was Fun
Running simulations of meteor swarm’s evolution
Władysław M. Turski 7
It Was a Mistake
I thought the computer was for computing
(Just as Aristotle thought brain was for cooling blood)
Władysław M. Turski 8
But it took me long time to see the light
First I thought computers were too difficult to program
Władysław M. Turski 9
But it took me long time to see the light
So I wrote compilers
KLIPA, ALGOL-60
Władysław M. Turski 10
But it took me long time to see the light
Then I thought running programs on computers was too messy
Władysław M. Turski 11
But it took me long time to see the light
So I wrote an operating system
SODA
Władysław M. Turski 12
But it took me long time to see the light
Then I thought something was
amiss with the data
Władysław M. Turski 13
But it took me long time to see the light
So I played with data structuresSo I played with data structures
Władysław M. Turski 14
And I’ve sinned
For five long years I worked in the industry.
Directing the transfer of OS/360 to Polish-made RIAD computers.
Władysław M. Turski 15
Much later (1983, Dublin)
Responding to an invited paper I said:
Fred Brooks is a hard act to follow: I know, I’ve lifted the OS/360 !
Władysław M. Turski 16
Then I thought I saw the light
Computer is for executing programs!
Władysław M. Turski 17
Programming methodology
Structured programming
(programming structures)
GOTO considered harmful
Władysław M. Turski 18
Programming methodology
Step-wise refinement
Correctness
Władysław M. Turski 19
Programming methodology
Correctness, yes but what does it mean?
With respect to ...
Władysław M. Turski 20
Programming methodology
Correctness wrt formal specification
- „faultily faultless”
- clear, but useless
Władysław M. Turski 21
Programming methodology
Specs
Reality Program
Correctness?
Ill-defined relationship
Władysław M. Turski 22
Programming methodology
Theory
Model Model
SatisfactionSatisfaction
Well-defined relationship
Władysław M. Turski 23
I was sure I saw the light
It all reduces to theories.
A good theory leads to good programs and satisfied customers.
A poor theory hinders construction of good programs and makes customers weep.
Władysław M. Turski 24
I was sure I saw the light
Build theories of small domains, make program-models for them, then combine theories and programs shall jell together!
Władysław M. Turski 25
I was sure I saw the light
But it ain’t necessarily so!
• combining theories is difficult (and not always possible)
• combining implementations is (nearly always) messy and often dangerous
Władysław M. Turski 26
I was sure I saw the light
I thought that perhaps combinig little actions into larger systems should be done entirely without any notion of „time”, i.e even without any predetermined notion of „follows”.
Władysław M. Turski 27
I was sure I saw the light
Thus the concept of doubly guarded command
)( rearactionfront
was born and (partially) explored
Władysław M. Turski 28
But while I followed the light
The world was changing
became the soup du jour
Władysław M. Turski 29
In the new brave world
Computers,
becoming ever more numerous (and physically smaller),
disappear behind services they provide.
Władysław M. Turski 30
In the new brave world
Computation has to be correct.
Service has to be dependable.
Władysław M. Turski 31
In the new brave world
The mathematical, strict
criterion of correctness
is being replaced
by a behavioural, loose
criterion of dependability.
Władysław M. Turski 32
In the new brave world
All actors (people, machines and systems thereof) occasionally malfunction.
This is not only accepted, but also expected.
Władysław M. Turski 33
In the new brave world
No-one really complains that Windows is prone to crash,
only that it crashes at an inopportune moment.
Władysław M. Turski 34
In the new brave world
Like no-one complains
that tires occasionally go flat,
only that it happens when we are particularly short of time.
Władysław M. Turski 35
In the new brave world
No-one is to blame and the guilt is subtly shifted to the „user”:
If I only didn’t drive there, if I only didn’t press these two keys together!
Or was it three?
Władysław M. Turski 36
In the new brave world
Every hour more users than babies are born.
Each user soon becomes an expert at their specific fingering.
Such an expertise quickly becomes an article of faith and an academic discipline.
Władysław M. Turski 37
In the new brave world
Particular fingerings turned articles of faith cum academic disciplines are invested.
Władysław M. Turski 38
An aside, from Webster’s:
invest
- to array in the symbols of office and honor
- to furnish with power or authority
- to surround with troops or ships so as to prevent escape or entry
- to commit money in order to earn a financial return
Władysław M. Turski 39
In the new brave world
In the ecumenical primordial ocean of Internet,
growth is primarily by accretion of external grime.
Reason is yet to be born: while no-one convinces, many try to catch.
Władysław M. Turski 40
In the new brave world
Correctness, rigour, mathematics of computing is pushed into an ever smaller and darker corner labelled
„special concerns”.
Władysław M. Turski 41
I have the feeling that
Having witnessed the birth of computing science,
I am witnessing its demise.
Władysław M. Turski 42
While it lasted
It certainly was fun!
43 Władysław M. Turski
?