Transcript Functional Paleontology: The Evolution of User

Functional Paleontology:
The Evolution of User-Visible
System Services
Authors: Antón, A. I. and Potts, C.
Presented by Shan Li
CISC 864: Mining Software Engineering Data
Outline

Discussion of the paper

Contributions of the paper



The approach used in the paper
Personal Perspectives



Purposes, examples, results, significance
Pleasant points in the paper
Unpleasant points in the paper
Conclusion & Questions
Before the Discussion
Requirements
Forward
Engineering
Designs
Codes
Reverse
Engineering
What is needed to be known?

The balance between in forward engineering
and reverse engineering is about the
requirements which emphasized in
forwarding engineering, while de-emphasized
in reverse engineering.
The Research Purpose
To redress the balance
To discuss the implication of
general patterns of functional
evolution for
forward engineering
To achieve the general patterns
of functional evolution
Research Methods


To restrict the study to the functionality
evolution
To propose an approach: functional
Paleontology



Paleontology: the study of the forms of life existing in
prehistoric times, as represented by the fossils of plans,
animals and other organisms
Analyzing fossil record of service evolution for a single
system
Services: collections of functions related to purposes or
mode of use
Research Methods

To characterize evolution in systems, based on
Shaw’s classification scheme for software
engineering research
Research Methods

To investigate into fossil record in an example



The telephony services available to domestic
subscribers over a 50-year period
Specifically, services contained in the call guide of
the Atlanta telephone directories for the years
1950-1999
To study the profile of services benefits and
burdens in a system’s evolution
Research Methods

To develop a conceptual base

Functional morphology:




Overall profile of benefits and burdens exhibited by a
system during its evolution
Benefits: system-use outcomes which meet a
beneficiary’s goals
Burdens: which undermine these benefits
Functional evolution:

Vocabulary of Evolutionary types
Research Results

The results of the telephony services
evolution






Categories of Benefits
Categories of Burdens
Service Epochs and Expansions
Punctuated Evolution Pattern
Periodic Retrenchment
Functional Decentralization
Research Results
Benefits profile over 50 years
From 1950-1999
Research Results
Burden profile over 50 years
From 1950-1999
Research Results

The evolution of telephony services suggests
two general trends:

Benefits for the actor beneficiary precede benefits
for others


Services initially benefits the people who actually
perform tasks
Object-level benefits precede meta-level benefits


Object-level: the creation and transfer of information
Meta-level: how the system is manipulating information
Research Results


Customization features may be relatively
unimportant in most systems. Let’s put it
other way, user customization may be
significantly limited by the legacy of core
design decisions
The approach is relevant to the management
of long-life software systems.
Future Research

Feasibility Questions





Is it possible to relate the functional morphology of
the system in question to the software
architecture?
Characterization Questions
Method Questions
Generalization Questions
Selection Questions
Personal Perspective

Pleasant Points




The paper has excellent structure and describes
their research method from the general level to
case study level; and explains their discoveries
from case study level to the general level
The paper defines every terms used.
The paper restricts the study into an aspect of a
system
The paper discusses the disadvantages of the
results based on the limited data
Personal Perspective

Unpleasant Points

The paper borrows many terms from other
disciplines, biology, engineering, linguistics.


Such as paleontology, morphology, langue..
The paper does not emphasize the implication of
the results for forward engineering of software
engineering, which is one of the purpose of the
paper.
Conclusion

Contribution of the paper



Proposed a new methodology, generalized their
approach by applied to most systems dominated
by “dynamic information systems problem frame”.
The approach is therefore relevant to the
management of long-life software systems
The research is not only for software systems, but
also for others
Thank You & Questions