> nt of the cases, or even 100% (why should I call get_is_molecule() > if I know that is going to be true all the time ? ). It is far more > difficult, but far more useful, to agree of what we expect to get > after a PDB, and how we are going to organize our structures in IMP > into hierarchies. You need some way of traversing the resulting hierarchy and telling what you are looking at. Say read_pdb returns a Hierarchy. You then have to go through the children of the node and determine which is a protein and which is a ligand if you want to do much useful with it. The docs for the read_pdb function can't possibly specify the answer. If you are writing code that knows the pdb that is being loaded, then you can hard code it, but that is not always possible (say you are writing something like modpipe or an application to handle fitting into em maps).
Or you load a protein and decide you want to define subdomains. You then modify the hierarchy to have a node corresponding to the structure. Some code which alter traverses the hierarchy (such as code to build a simplified representation), may need to know when it has gotten to the atoms or residues (it can't assume a constant structure).
> Together with Keren, I can give suggestions for low resolution > structures. Looking forwards to it.