[CraigS]

Originally Posted by Robh

Interesting. A chaos-based simulation has to evolve with dynamical interactions on many levels. In fact, any model that ignores key influences is going to lead to misleading results.
This is all going to take an extraordinary amount of computer memory and processing capability considering the N-body description alone involves around 10^12 stars over an enormous amount of simulated time. With so many interconnected processes, one would have to question the reliability of any conclusions deduced by current simulations. The effects of dark matter is yet surely uncertain.
My main bone of contention is that a simulation that reproduces what we see today is not necessarily based on an accurate hypothesis. If B is true and A implies B, it does not mean A is true. For example: "the numbers 3 and 4 are prime" implies "3 is prime", the latter is true but the former is not.
A simulation must not only show what we see to be true but predict something that is not yet seen to be true to add credibility to its assumptions. Example, some unknown follow on from newer galaxies through to older ones. A process akin to that of a classical scientific theory.

Regards, Rob

Hmm … looks like you read the 'Introduction', Rob.
Its a good piece, I find it to be fairly balanced ...

AS Pfenniger says:

Since physical theories try to build an isomorphism between Nature and a subset of mathematics, it follows that no “theory of everything” (TOE) can summarize with provable derived theorems the implicit complexity of its content.

As an aside: Its kind of amusing too, that I started out this thread with a bone of contention … now Rob has one …

I guess Rob's bone might be the incorporation of Dark Matter ? Pfenniger talks about three types .. neutrinos, axions and CDM. He says:

The often adopted collisionless property of cold dark matter is just an assumption that may be acceptable in present day galaxy models, but may turn out to be invalid during perturbation sensitive events like pancake or filamentary collapses.

The presense or absence of DM is never questioned. Why would any self-respecting scientist question this aspect, anyway ?

As for predictions: models are the computer-age equivalent analogy to classical theories so, yes, they should be expected to make predictions which can be tested. In the case of Galaxy simulation, there is a lot that goes into them though. Dark Matter is only one part of it all. The sensitivity of the predictions compared against each hypotheses would also seem to be a testable outcome, though, eh ?

Robert Grand's predictions might be break-up of spiral arms and that leading and trailing edge stars migrate towards and away from the core of the galaxy, eh ? Their dependence, or sensitivity to the hypotheses (or assumptions), would be a very interesting topic. Perhaps we should look for this aspect in future reports about computer model predictions ?

Also, surely if our hypotheses about DM were completely off-track, then we would never end up with the complex, highly recognisable shapes which they do produce ?

Interesting (& complex ! )

Cheers