# Quantitative test of diffusion on 2D surfaces
/*
Results are time-step dependent, at least for long time steps.  With dt = 10, profile
is a perfect Gaussian, because no allowance is made for diffusion between t1l and t1u,
which is the correct answer.  For shorter time steps, this diffusion occurs and the
concentration on the sum of t1l and t1u increases.  This is physically correct.  So,
Smoldyn appears to be working correctly here.
*/


define SEED 1

graphics opengl
random_seed SEED

dim 2
boundaries x 0 100
boundaries y 0 100

species red

difc red(all) 10

color red(all) red
display_size red(all) 5

time_start 0
time_stop 9.99
time_step 10

frame_thickness 0

start_surface surf1
color both black
#panel rect +y 0 50 100 r2

panel tri 15 24.0192  30 50  t1l
panel tri 15 75.9808  30 50 t1u
panel tri 30 50  60 50  t2
panel tri 60 50  100 50  t3

neighbors t1l t2 t1u
neighbors t1u t2 t1l
neighbors t2 t1u t3 t1l
neighbors t3 t2
end_surface

surface_mol 1000000 red(up) surf1 tri t2 50 50

text_display red(all)

output_files FILEROOTout.txt
cmd A molcountspace red(all) x 30.0001 100 140 0 100 0 FILEROOTout.txt
cmd A molcountspace red(all) x 15 29.9999 60 0 100 0 FILEROOTout.txt

end_file

