> if cgal can be installed painlessly, i.e., a normal person > (non-bioinformatics person or non-computer scientist) can install it > with few mouse-clicks or commands (<5) on major platforms (linux, mac, > windows), i think nothing would speak against it. also the lib should > not mess up the harddisk. I have only installed it from the command line, but that is a matter of typing "./install_cgal --prefix=where_to_install". There are binaries for windows (and I'll check on macs).
> if imp depends on libs that are hard to > install it will die rather sooner than later. Agreed. Nothing should be harder than "./configure; make install" or running a gui installer (which I think is more of a pain :-) But then again, I can't think of why any library should be harder than that. But some are.
> but otherwise, i think we > all want to save time wherever possible. > one possibly ignorant question: andrej got along without fancy libs. > knowing him, i presume he stole the code from somebody else. can't we > simply do the same? the code was pretty fast...
:-) The modeller code doesn't work because it assumes equal sized particles. He also claims he put a fair bit of effort into optimizing it. I don't really feel like spending my time that way. I have looked for good public domain code to steal, but haven't found any.