Like how accurate are these models? Could these be software or hardware errors? I mean, are these running on 1970 terminals in Fortran or is there real development going on? I imagine it's super hard to measure such tiny disrepancies in the cosmos when we are failing constantly in common modern software.
back
1 comments
I mean, if it’s Fortran that’s been around since the 70s, it’s probably been around long enough that it is both better vetted for accuracy, and not burdened by “real development” affecting some static mathematical principals that the code is computing….
Valid argument, my main point though is how do we know whether these are not errors in the model generation and not in the universe acting in weird ways? It's not that you can make tests against the universe, you have to always rely on some algorithms giving you the results, which could be wrong and every institute could be using the same ones from the same vendor. Where are the data and what is the software is my worry.