UNIVERSITY OF HERTFORDSHIRE
		COMPUTER SCIENCE RESEARCH COLLOQUIUM

			     presents


		    "The Impact of Bad Smells on Code"

			     Min Zhang
	(School of Computer Science, University of Hertfordshire, UK)


		     28 April 2010  (Wednesday)
		      Lecture Theatre E351
		   Hatfield, College Lane Campus
			      3 - 4 pm


		   Everyone is Welcome to Attend

		   Refreshments will be available



Abstract:

This presentation is on my PhD research into the relationship between
Code Bad Smells and software faults. Fowler et al. state that Code Bad
Smells can give "indications that there is trouble [in code] that can
be solved by a refactoring". However, my review of current software
engineering literature on Code Bad Smells shows that little empirical
evidence has been provided to support Fowler et al.'s claims that Code
Bad Smells cause detrimental effects on software. Fenton and Pfleeger
suggest that software faults are an important indicator of software
quality. As a consequence, software problems caused by Code Bad Smells
are likely to be indicated by software faults. Hence, it is reasonable
to expect that Code Bad Smells can cause faults in software. My PhD
research empirically investigates the impact of five of Fowler et al.'s
Code Bad Smells, i.e., the Data Clumps, Switch Statements, Speculative
Generality, Message Chains, and Middle Man, in terms of their relation
to faults.

My research captured the relationship between Code Bad Smells and
faults using source data from the Eclipse and Apache Commons open
source projects. My results from both of these datasets suggest that
source code containing only one of the five investigated Code Bad
Smells is less fault prone than source code which does not contain any
of these Code Bad Smells. This indicates that the existence of only one
of these Code Bad Smells in source code is not likely to make it fault
prone. However, my study using Eclipse data suggests that all source
code samples associated with most faults contain the Message Chains Bad
Smell. In addition, Message Chains are associated with a high number of
Algorithm/Method faults and the other Code Bad Smells are associated
with Internal Interface faults. Since the Algorithm/Method fault can
cause bigger problems in software than the Internal Interface fault, it
leads me to conclude that the Message Chains Bad Smell deserves further
attention.


-
Hertfordshire Computer Science Research Colloquium
http://homepages.stca.herts.ac.uk/~nehaniv/colloq