Title: School of Information and Computer Science
1Beyond Programming The Philosophy of Science
Author
Presenters
Ping Chen Chris Van der Westhuizen
2Introduction
- There is a tension that exist in the software
development process - Identification of the requirements to the
developing programs and models - Similar tensions exist in the sciences
- Describe the real world with models and theories
3Introduction
- There is a tension that exist in the software
development process - Identification of the requirements to the
developing programs and models - Similar tensions exist in the sciences
- Describe the real world with models and theories
Goal Resolve this tension
4Why Philosophy of the Science?
- A kind of meta-science
- Examines the nature of science
- Help determine
- What is knowable
- What can be represented formally
5Four Areas of Philosophy
- Logic evaluation of an argument
- Metaphysics determine the fundamental types of
things and how to describe them - Epistemology determine what knowledge is and
how it is gained - Value theory how to establish norms
6Logical Positivism
- Knowledge based on experience
- Able to model natural phenomena
- Use of deduction to explain events from laws and
initial conditions - Unified science
7Problem of Logical Positivism
- Logical positivism laws are based on induction
Will always get fed at 9 AM
Fed at 9 AM everyday for the past few months
8Problem of Logical Positivism
- Logical positivism laws are based on induction
Will always get fed at 9 AM
Fed at 9 AM everyday for the past few months
Christmas at 9 AM
9Popper
- Need to examine current theories
- First make conjectures then try to disprove them
- Severe tests leads to greater confidence
- Failed theories lead to new theories
- Progress comes from bold conjectures that are not
refuted
10Campbell
- Evolutionary epistemology
- 10 process levels
- From non-mnemonic problem solving to science
- Science
- Knowledge is testable
- Have a selective system to weed out conjectures
- Designed as objective as possible
11Campbell (cont.)
- Criticizes Poppers view of the natural selection
of theories - Proposes blind-variation-and-selective-retention
process - Mechanisms for introducing variation
- Consistent selection process
- Mechanisms for preserving and/or propagating the
selected variation
12Quine
- Both Popper and Campbell recognizes that
observations are not objective - Cannot use our experiences to either confirm or
falsify hypotheses - Two completely different theories may still be
empirically equivalent
13Empiricist Account of Science
- Emerged by the late 1930s
- Based on 3 assumptions
- Naïve realism
- Universal scientific language
- Correspondence of truth
14Relation to Software Development
- Logical positivists methods for knowing what is
true and how to build on that knowledge - Popper only know what is wrong
- Campbell progress be evolutionary
- Quine unable to have objective observations
- Specification are underdetermined
15Kuhn Science in a Historical Context
- Kuhn detected five distinct stages in the growth
and evolution of a scientific discipline - Immature science
- Normal science
- Stages of crisis
- Revolution
- Resolution
16Stages of Evolution
3. Stage Of crisis
1. Immature Science
2. Normal Science
4. Revolution
5. Resolution
17Normal Science
- Operates within a paradigm that establishes a
framework for practice of scientific discipline - Normal-scientific research is aimed at the
articulation of phenomena and theories that the
paradigm already supplies
18Paradigm Shift
- A discovery like x-rays necessitates a paradigm
change for a segment of the community - Identification of anomalies
- Emergence of new discoveries
- Transition from a paradigm in crisis involves a
reconstruction of the field from new fundamentals
19Paradigm Shift
20Paradigm Shift
Paradigms do not coexist but rather move from one
to another
21Scientific advancement?
- Does the acceptance of improving scientific
efficiency imply that science constantly
approaches some goal? - Kuhn thinks not
- The paradigm shifting process occurs as
biological evolution does, without benefit of a
set goal
22Lakatos
- Differed from Kuhn
- Believed science is seldom dominated by a single
paradigm - Believed that new theories replaced old theories
while retaining some important features - Introduced idea of research programme
23Lakatos Research Programme
24Lakatos Research Programme
- Examples
- Copernican programme requires acceptance that
planets revolve around the sun - By rejecting this, Tycho Brahe opted out of the
programme
25The Structure of Science
- Abrahamsen provides a specialization hierarchy
for the scientific disciplines - Four-level hierarchy
Cultural Product Domains
Behavioral Sciences
Biological Sciences
Physical Sciences
26The Structure of Science
- Campbell introduces a fish-scale model of
omniscience - Knowledge is organized in an overlapping
structure (ideal)
27The Structure of Science
- Campbell also pointed out the current structure
in disciplines - Interdisciplinary gaps (not ideal)
28Relation to Tension of Software Process
- Formal specifications have no validity without
the theories that provide their context - Science world view comes from interpretation of
reality - Software world view comes from interpretation of
a need
29Looking Ahead
- The rigor and critical attitude of science are
necessary for software development - but they are not sufficient
- Our criterion for success will be how well this
science serves us
30Discussion Questions
- How does understanding the tension that exists in
physical sciences help us as computer scientists? - What is the received view? (fig. 2.1). How does
it fit in? - What is difference between evolutionary models
proposed by Campbell and Popper? - Can criticism be objective? (according to Popper
it is)
31Discussion Questions
- Can science be unified in one theoretical
framework? (positivist theory) - What are some software paradigms that exist? Have
existed? - What philosophical perspective do you agree with?
- Khuns single paradigm
- Lakatos research programme