This preview shows page 1. Sign up to view the full content.
Unformatted text preview: key responses of effort, interval,
quality. We
present three such indices, two dealing with quality and and
present three such indices, two dealing with quality
one with effort, which are discussed more thoroughly
one with effort, which are discussed more thoroughly in in
already been defined as Sections 5.3
eady been defined as Sections 5.3 and 5.4.
and 5.4.
commentary source lines,
Predictors of the number faults that will have to be
mentary source lines,
Predictors of the number of of faults that will have to be
ll files f belonging to m. fixed in module m (and, thus, of the quality response) in a
les f belonging to m. fixed in module m (and, of faults the quality response) in ain module m
Fault Potential 1. The number thus, of that will have to be ﬁxed
te in Section 5, extensive future interval of time, taken from [20], include the
in f ture i terval effects are dampened [20 , in
n Section 5, extensive theufuture. nChange of time, taken fromover ]timeclude the
d software ªcomplexityº weighted time damp model
oftware ªcomplexityº
EICK ET AL DOES CODE model X
ly correlated with xgv in weighted .:time dampDECAY? ASSESSING THE EVIDENCE FROM CHANGE MANAGEMENT DATA
orrelated with xgvwith
ph
m; t
I
X eÀtÀhei
fectively synonymous in
ÀtÀhei
p to
mless than m; hei <
estimatedhbe ; t
c e>one, as
in t oure data). Statistical 5.1 Temporal Behavior
I
vely synonymous with
R
c > m; hei
<
analyses of the models appear inSectiont 5.3.
; m
RThe CDI FILES of (3) meas
Â loge
ehh
; m hiv
Â logfaults
that willhiv
to be ﬁxedhow many code units (files
4.2.6 Effort
Fault Potential II. The number of ehh ; m have ; m
in a module • • implement it. An increase
a software unit m as the and the generalized linear model
One may construct predictors of the effort (personhours) symptomatic of decay, as d
X
in
ines. The unitof eqi the a and the generalized linear model> code (when beta is <1)
use m as in the future. Faults are less likely in older
oftware
1fc e b g f t eqi
m e
S
requirepqvw
m; tl mentÂa change my Âiting r; gression
d to i p e P
Fig. 3 shows that span
Section 4.2.5.
X
cPÁ effort as functions ofm
. The use of eqi in a
equations for measurements of
1fc e> mg Â eqi
factors
Sunder study. There, we dis
;
pqvw
m; t P Â
ility of the ages of the lines
10
that affect effort. A sample form for such and eqi is
ion[19] in Fig. 2 shows the where Á is the entire set of deltas up to time t a regression time an MR touches more
4.2.5.
cPÁ
w
relationship Section 4.2.4
of the ages of the lines , in which each point corre
Effortdiscussed in is:
Model:
personhours , I , and P
10
e. The files are represented where Á predictors of theand where up to time t and are
eqi is
is the statistical analysis.
estimated usingentire set of deltas X
coordinate is time represen
19] in Fig. 2 shows the
ged for confidentiality) and
1 , e f , and are
EFF
cSection a4.2.4 and a2
a0 1 FILES
c where fc ,> gjf j
Both of these indices illuminate change as theI primaryP
ycoordinate is one when m
he files are represented discussed in
f
resented by colored lines
agent creating statistical analysis.
estimated usingfaults. (Even though faults do not arise zero otherwise. Three loca
6
for these colored lines are
of confidentiality) and
a3 ADD
c a4 DEL
cabsence of other
spontaneously, this is not a tautology: The
and
Both of these indices illuminate change as the primary [22] for introduction a
fnted by colored lines of
characters in the lines
terms such as faults. complexity DEV : informative.)
and (Even a6 is highly
top
agent creatingsize a5 INT
c though
cfaults do not arise plot. These smooths
interpreted as lookingare a
at
these colored lines
averages, where the weigh
The indices depict differing tautology:effects. In (4), the
temporal The
Here, jf j denotes the size a
f.
tance. The color lines line spontaneously, this is notin NCSL of the file absence of other
of...
View
Full
Document
This document was uploaded on 03/16/2014 for the course EE 360f at University of Texas.
 Spring '08
 Staff

Click to edit the document details