home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.umcs.maine.edu
/
2015-02-07.ftp.umcs.maine.edu.tar
/
ftp.umcs.maine.edu
/
pub
/
WISR
/
wisr6
/
proceedings
/
ascii
/
matsumoto.ascii
< prev
next >
Wrap
Text File
|
1993-10-19
|
10KB
|
237 lines
A Nation-Wide Evaluation of Software Components Reusability
Masao J. Matsumoto
C&C Software Development Group
NEC Corporation
Shibaura 2-11-5, Minato
Tokyo 108 Japan
Tel: +81.1.5476.1090
Fax: +81.1.5476.1095
Email: matumoto@ccs.mt.nec.co.jp
Atsuko Ishida
Institute of Advanced Business Systems
Hitachi, Ltd.
890 Kashimada, Saiwai, Kawasaki
Kanagawa 211 Japan
Tel: +81.44.549.1530
Fax: +81.44.549.1528
Email: ishida@iabs.hitachi.co.jp
Abstract
Much discipline is needed in order to enable software engineers to develop versatile reusable
components. There is need for a set of criteria which enable software builders to evaluate software
component reusability both quantitatively and qualitatively across domains. This position paper
describes the evaluation criteria, proposed methods, and related considerations, surveyed in a
study circle with participation from experts in both industry and academia. However, since
reuse disciplines defined through this time survey have not been properly verified, followup
studies are needed.
Keywords: Reusability evaluation criteria, Software quality characteristics, Reuse discipline.
Workshop Goals: Learn and exchange information on recent reuse research results.
Working Groups: Standard guidelines for reuse, Quantitative evaluation
Matsumoto- 1
1 Background
As a member of NEC's Software Engineering Central Group, M.J. Matsumoto has facilitated the
development and exploitation of software reuse libraries in the management information systems
domain. He has undertaken the role of architect in the development of a software engineering
environment which integratesreuse supports for office server systems during 1980's. As chair of
numerous software quality research circles at the Japan Union of Scientists and Engineers, he has
promoted several research thrust programs (including software reuse), attended by delegates from
many leading industries and academia. Ms. A. Ishida from Hitachi Ltd. has headed the software
reuse-related study teams. These joint study activities provide not only an opportunity for members
to have insight on reuse issues, butalso a chance to explore possibilities for forming a nation-wide
consensus on the discipline of reuse engineering.
2 Position
Not all reusable components are reused at the same frequency. Some are very frequently reused,
while the others somewhat less. What does this mean? For a given component, no reuse for a
certain time period means perhaps that the component does not have enough versatility, with the
more frequently reused component having more versatility. The reuse frequency is an index that
can be used for evaluating the value ofa reusable component, though the frequency itself does not
mean the amount of money savings whichthe reuser can obtain at each reuse time.
If it is true that each component has its own reuse frequency, where does this frequency come from?
Is it inherent to the component itself or to the characteristics of the reusers? Not enough surveys
have been made for making judgements on what the ma jor factors are thataffect this frequency
distribution.
There seem to be several factors which prevent software engineers from reusing software compo-
nents, some of a managerial and some of a technical nature. The technical factors must be at
least three-fold: factors pertaining to the component itself, factors pertaining to the reusers of the
component, and factors pertaining to the environment in which the component is reused. This time
survey focused on the component factor, while the other factors need further preliminary investi-
gation before surveys can be started. The primary aim of this time survey was to find relationships
between characteristics and reusability(reuse frequency) for a given component.
3 Comparison
3.1 Evaluation Criteria
There are few well-defined sets of criteria for evaluating software component reusability that have
been commonly agreed upon in the reuse community. One possible set could be the 21 character-
istics defined in ISO/IEC 9126 Annex A [1]. This set was originaly defined for evaluating software
quality from viewpoints such as functionality, reliability, usability, efficiency, maintainability and
portability. Since the ISO 9126 is a criteria which is used for evaluating software quality, it should
be helpful for finding relationships between the reusabilityof given components when reviewed from
the viewpoint of quality. This is the hypothesis we have taken as our position.
Matsumoto- 2
HO: The ISO 9126 could be used as a criterion for evaluating the reusability of software
components.
The REBOOT criteria for software reusability might be another candidate which could be used for
this evaluation purpose [2]. Metrics defined in the REBOOT might be of help in measuring each
criterion quantitatively. The REBOOT criteria, however, have not yet widely been recognized as
standards at the international level.
3.2 Proposed Method to Quantify Reusability
A set of reusable components was carefully examined, each component of which was reused fre-
quently and with considerably high reputation. 21 components were sampled from this set. Another
set of seldomly reused "reusable" components was examined, and five components were sampled
from this set. Several member companies offered sample reusable components under these two
distinctively defined categories: frequently reused and seldomly reused [3].
For each of these two distinctive sets, we gave scores to each of the components using five ranking
methods, one for each of the ISO 9126 sub-characteristics. Comparing the component scores for
these two sets, we have tried to find properties or quality characteristics which each category of
components commonly holds.
3.3 Considerations
We found that the characteristics we must pay attention to in terms of factors which affect com-
ponent reusability may varydep ending on the software size, execution environment, development
methods, and machine type. Consequently,guidelines for better reuse design should be established
for each of these circumstances. We are however able to point out some disciplines, for example,
where there is a big gap between two frequency categories in terms of functionality, reliability,
usability and efficiency.
Accuracy, connectivity and understandability are thought to be factors independent of reusable
component circumstances. On the other hand, characteristics which are thought to b einmp ortant
vary largely from one domain to another. For example, understandability is thought to be important
in the downsizing systems domain, while it is not in the embedded or communication systems
domain. This example shows that further studies must be undertaken for defining disciplines which
can be commonly used across domains.
References
[1] "International Standard Information Technology - Software Product Evaluation - Quality Char-
acteristics and Guidelines for Their Use, First Edition," Tech. Rep. ISO/IEC 9126, ISO/IEC,
1991.
[2] A. de Mora etal, "A Metrics Approach to the Software Reuse Problem," in Proceedings of the
Third European Conference on Software Quality, 1992.
[3] Interim Report: Software Reusability in The 8th Year-Round Workshops on Software Quality
Studies, pp. 133-141. Japan Union of Scientists and Engineers, 1992. (in Japanese).
Matsumoto- 3
4 Biography
Masao J. Matsumoto joined NEC, Tokyo, Japan, in 1963. His responsibilities on the NEC
Software Engineering Laboratory included modeling, developing, and exploiting integrated soft-
ware development environments for and with reusable components under the SEA/I project. He
had helped formalize and simplify software reuse libraries under the p ost STEPS, SSPL and SPar
projects. He currently works in the C&C Software Development Group, NEC, and participates in
corporate software engineering programs. He is developing software engineering curricula encom-
passing reusability, quality and productivity for the NEC Institute of Technology Education. He
also chairs the Software Quality Study Programs at the Japan Union of Scientists and Engineers. He
gives lectures on software engineeringat Yamanashi University,School of Engineering and Tsukuba
University,Graduate School. Dr. Matsumoto earned his bachelor's degree in Mathematics Science
at Waseda University, and his Ph.D.degree in Information Engineering from Kyushu University.
Ms. Atsuko Ishida joined Hitachi, Ltd. in 1972 after graduating from Tokyo University, Faculty
of Sciencei, earning earning her bachelor's degree in Mathematics. She has led several successful
projects in software development and engineering. Ms. Ishida has also headed software reusability
study groups at the JUSE.
Matsumoto- 4