I think it would be really nice if we could extract distributions of compute times for various tasks, in particular ancestral state reconstruction, so we can make better decisions about how we downsample. I think SCons might actually record some of this information in its .sconsign.dblite file, so we'd really just have to extract it and do some aggregation to compute some distributions. This probably isn't that high priority a thing. But if we're looking for things to do and the data is in the sconsign already, it might be worthwhile.
I think it would be really nice if we could extract distributions of compute times for various tasks, in particular ancestral state reconstruction, so we can make better decisions about how we downsample. I think SCons might actually record some of this information in its
.sconsign.dblite
file, so we'd really just have to extract it and do some aggregation to compute some distributions. This probably isn't that high priority a thing. But if we're looking for things to do and the data is in the sconsign already, it might be worthwhile.