On Wed, Aug 4, 2010 at 8:51 AM, Christian D. Ott cott@tapir.caltech.edu wrote:
Hi,
On Wed, Aug 04, 2010 at 08:32:56AM -0400, Roland Haas wrote:
Hello all,
On the 'downside' this would introduce a dependency of HydroBase on TmunuBase, even with a parameter to turn it off. I don't think anyone would not use HydroBase without TmunuBase in the foreseeable future though.
Wouldn't doing simulations using the Cowling approximation not require TmunuBase?
What are the opinions about the proposal to move setting Tmunu from GRHydro to HydroBase?
I have slight preference for not having HydroBase setting Tmunu since doing so adds code to HydroBase making it a more active thorn and not just a thorn that defines a common interface for other thorns.
one thing to keep in mind is that there are multiple definitions of the 3-velocity out there. HydroBase would have to 'know' which on a given hydro thorn is using to be able to set up Tmunu.
To avoid confusion, the velocity in HydroBase must always have the same definition, independent of the evolution code. This is currently the same convention as Whisky/GRHydro.
If we keep the Tmunu calculation out of HydroBase, then we should still have it in one central place, and not repeat it in every hydro evolution code. We could introduce a new thorn CalcTmunu that inherits from both HydroBase (reading its variables) and TmunuBase (setting Tmunu). Personally, I find this more complex than necessary, and there is the danger that people will forget to activate this thorn, so I would instead include this functionality directly in HydroBase. For Cowling, the TmunuBase variables won't have storage, and HydroBase can detect this.
-erik