# 2021.01.03 JH

# See http://www.medicine.mcgill.ca/epidemiology/hanley/BridgeOfLife/

# This R function replaces the Java app JH wrote
# in 2008, but that is becoming more difficult to
# run on platforms that have added security measures.

# STEP 1: begin by sourcing the R function
 
source("http://www.biostat.mcgill.ca/hanley/BridgeOfLife/BoLfunction.R")


# STEP 2: decide which lifetable you wish to animate

# options:

# 1 Sweden 1751-1851 Females ['COHORT' lifetable (LT), born 1751 ]

# 2 UK 1870-1880 Males  ['CURRENT'/'PERIOD' LT, using death rates from 1870s]

# 3 France 1895 Males     ['CURRENT'/'PERIOD' LT, using death rates in 1895]
# 4 France 1895 Males     ['COHORT' LT, cohort born in 1895]
# 5 France 1895 Females   ['COHORT' LT, cohort born in 1895]
#            (the year 1895 was chosen to contrast CURRENT vs COHORT)
#            (those born that year were 19 when WW I began)

# 6 Switzerland 1895 Males   ['cohort' LT, cohort born in 1895]
# 7 Switzerland 1895 Females ['cohort' LT, cohort born in 1895]

# 8 Canada 2000-2003 ['current'/'period' LT, using death rates in 2000-2002]

# STEP 3: After making the plot window as large as possible, call

BoLfunction(option = 1,        # or replace the 1 by another option 2-8 
            n.cohorts=1110,    # change as desired
            DELAY=0.1)         # change as desired

# NOTES

# When you call this function to animate the option,
# You might want to experiment with the DELAY parameter
# that indicates how many seconds delay between frames

# On JH's MacBook Pro, 0.1 seconds works well,
# On faster machines/slower machines, you might want to
# change it to say 0.05/0.5 seconds

# the FIRST TIME you try it, you might want to play 
# quite slowly (e.g. 5 seconds/frame) just to see HOW IT WORKS:

# ie 10 have their 0-th birthday (are born)
# the cohort immediately moves on 1 year, and each member either 
# survives the year (shown as a green dot) or dies during it.
# The latter 'events' are shown in red.
# every green dot counts as a full person year (PY) lived 
# every red dot counts as 1/2 a PY lived

# The next cohort appears, etc...

# the age-specific numbers of deaths during the year (the cycle)
# are also shown in red underneath the first panel.

# for the first 110 cycles, the app displays in green 
# in the second panel the number alive at each age
# and in red in the third panel the number of deaths at each age

# 110 years after the very first cohort was 'born' 
# (ie after the app reaches steady state) ...

# the app displays in green in the second panel the AVERAGE 
# (over the cycles thus far) number alive at each age.
# It is scaled up by 10, so it is out of 100 born.

# it also displays in red in the third panel the AVERAGE 
# (over the cycles thus far)  number who dies at each age.
# It is also scaled up by 10, so it is out of 100 born.

# It displays as red dots the rate of deaths per PY,
# and as a red curve the logs of these age-specific death rates
# (the log-linearity was first noticed by Gompertz and Edmonds).

# The n.cohorts parameter indicates how many cohorts 
# (each of size 10) will make their appearance.
# It also denotes the numbers of cycles

# Think of a cohort as you and 9 birthmates born at the same 
# time, and the cohort behind you as another  10, etc...
# (there isn't room to show your entire birtcohort at once)