Oops. I missed the part about it being static. Well, my answer still
holds if there's only 1 turtle in the simulation. Or if all the turtles are
the same age. Or if each turtle instance additionally stores a mutable
"ageOffset" which is added to the class age to determine each instances
particular age. The advantage of this last design is that we'd only need to
increment to single static age field to get all the turtles to age
simultaneously, rather than iterating through each turtle and aging them
individually (which may perhaps result in race conditions?)

   (I'm joking, of course)

   I don't know, in my opinion all those values might vary over time. If I
increase my suspension-of-disbelief threshold a bit, there are 3 fields
which conceivably would not vary over time (depending on how simplistic we
make our turtle-simulator app), and they're all sort of tied IMHO in terms
of plausibility.

   - Oliver

   The number of legs is one of the three way ties I was thinking of.
Average weight and longest recorded age are the others. Average weight might
not be the average weight of only the turtles in the simulator, but of all
turtles, past present and future. I.e. it might be derived from some model
that biologists have about turtles. E.g. the average weight of a hawksbill
hatchling is said to be 13.5 to 19.5 grams. It doesn't matter how many
turtles are in your simulator, nor what they weight; the average weight of a
hawksbill hatchling would remain constant regardless.

   Similarly for longest recorded age. If it is known that turtles of a
given species have a life expectensie distributed normally around 200 years
with a standard deviation of about 30 years, but one freak turtle was
discovered to have lived two thousand years due to special circumstances
(perhaps surviving in a special environment which no longer exists), we
could probably safely rule out the possibility of ever seeing another turle
live as long (at least with the same, if not greater, plausibility that we'd
rule out the possibility of a turtle ever losing a leg), and just hardcode
that 2000 years is the longest recorded age, regardless of what age the
turtles in our simulator would ever reach. Perhaps the longest recorded age
field might even be used as a sanity check, so that if the ages of any
turtles in the simulator reaches that, we know some sort of error occurred
somewhere.

   - Oliver