[FRIAM] nice quote

[Jon Zingale]

Since we are picking on aphorisms, I wish to add criticism to "all models
are wrong, some are useful".

To a great extent, the qualities of the thing being modelled matters. For
instance, natural numbers do have crisp, compact properties and can be
modelled by sets. To claim that models are never correct is to deny that
bisimulation ever exists between machines. It is fine, I suppose, as a
world view, but proving bisimulation between things also seems fine.

Weirder still is the case where things can model one another and yet not be
useful. While closed lambda calculi, turing machines and agent based models
can all model universal computers, limitations of various types can render
any particular model useless. Only partially with tongue-in-cheek, I am not
sure anyone has ever found Turing's machine to be useful for anything other
than getting a passing grade on a senior year project. In the case of ants,
do we really have to wait so long?

Alright, time to run away before my ears are George Boxed...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20241008/7d153a7e/attachment.html>