I write and maintain a rather large (100k+ LOC) codebase implementing various civil / mechanical engineering design codes.
I think the time has come to unitify the modules so I'm evaluationg options.
My questions are:
How does forallpeople perform when literally thousands of objects are present with units? As I will use the structural environment (with the addition of Kelvin, which strangely seems to be missing for this environment), I'd even consider kind of severing the SI system to exclude base units not needed to reduce the footprint.
Is any care to be taken when using forallpeople between modules and packages? Think the way the pint package requires to use the same UnitRegistry instance to be used within a calculations. This issue is not mentioned in the documentation so I guess this is not an issue?
Performance. At the start of tuplevector.py a todo states that performance could be massively enhanced if that module was rewritten. Is there any advance on this? Would that be a lot of work to do it?
Dear all,
I write and maintain a rather large (100k+ LOC) codebase implementing various civil / mechanical engineering design codes. I think the time has come to unitify the modules so I'm evaluationg options.
My questions are:
with the addition of Kelvin, which strangely seems to be missing for this environment), I'd even consider kind of severing the SI system to exclude base units not needed to reduce the footprint.Thank you in advance for your insights.