# r --vanilla <rInput.txt

# ==========================================================================
# ===== java.util data =====================================================
# ==========================================================================

javaUtilData <- read.table("v11data-javaUtil-edit.txt", header=TRUE, sep=" ")
names(javaUtilData)

# ===== Sums of columns ====================================================

sum(javaUtilData$SLOC)
sum(javaUtilData$N10t)
sum(javaUtilData$RC10t)
sum(javaUtilData$Coverage10)
sum(javaUtilData$N11t)
sum(javaUtilData$RC11t)
sum(javaUtilData$Coverage11)

# ===== Coverage ratios ====================================================

sum(javaUtilData$Coverage10) / sum(javaUtilData$SLOC)
sum(javaUtilData$Coverage11) / sum(javaUtilData$SLOC)

javaUtilData$Coverage10 / javaUtilData$SLOC
javaUtilData$Coverage11 / javaUtilData$SLOC

# ===== Means of columns ===================================================

mean(javaUtilData$SLOC)
mean(javaUtilData$N10t)
mean(javaUtilData$RC10t)
mean(javaUtilData$Coverage10)
mean(javaUtilData$N11t)
mean(javaUtilData$RC11t)
mean(javaUtilData$Coverage11)

# ===== Normality tests on column differences ==============================
# ===== p value is probability that data is normal =========================

shapiro.test(	javaUtilData$N10t -
		javaUtilData$N11t)
shapiro.test(	javaUtilData$RC10t -
		javaUtilData$RC11t)
shapiro.test(	javaUtilData$Coverage10 -
		javaUtilData$Coverage11)

# ===== t tests on columns =================================================
# ===== p value is probabililty that column difference has mean 0 ==========

t.test(		javaUtilData$N10t,
		javaUtilData$N11t,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)
t.test(		javaUtilData$RC10t,
		javaUtilData$RC11t,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)
t.test(		javaUtilData$Coverage10,
		javaUtilData$Coverage11,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)

# ===== Paired Wilcoxon tests on columns ===================================
# ===== p value is probabililty that column difference has mean 0 ==========

wilcox.test(	javaUtilData$N10t,
		javaUtilData$N11t,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)
wilcox.test(	javaUtilData$RC10t,
		javaUtilData$RC11t,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)
wilcox.test(	javaUtilData$Coverage10,
		javaUtilData$Coverage11,
 		alternative="two.sided", paired=TRUE, conf.level=0.95)

# ==========================================================================

# ==========================================================================
# ===== Apache data ========================================================
# ==========================================================================

apacheData <- read.table("v11data-apache-edit.txt", header=TRUE, sep=" ")
names(apacheData)

# ===== Sums of columns ====================================================

sum(apacheData$SLOC)
sum(apacheData$Nt)
sum(apacheData$RCt)
sum(apacheData$Coverage)

# ===== Means of columns ===================================================

mean(apacheData$SLOC)
mean(apacheData$Nt)
mean(apacheData$RCt)
mean(apacheData$Coverage)

# ===== Coverage ratio =====================================================

sum(apacheData$Coverage) / sum(apacheData$SLOC)

apacheData$Coverage / apacheData$SLOC