My preference would be for something implemented in Python/C. The project could start with a clean, extensible (it should be easy to add new molecules) and modular Python library and then start replacing some functions with C equivalents. I would like a library that doesn't use inheritance, or use it at less as possible (it makes too difficult to read code) which could only contain descriptions of molecules and maybe perform some basic operations (functions could be and are implemented in other libraries) Even when the library wolud be in Python, It should be easy to "plug" C functions into it so we would not loose in performance. It should also be as similar to chempy as possible so pymol could be easily ported to it. I don't know how well chempy fits with the avobe description, because, due of the lack of documentation, I only use it in a very basic way. Maybe It only needs minor changes and documentation to become the standar in chemical objects for Python.
It sounds an interesting project. I would probably join if something comes out Cheers Raúl >CDK is too Java-centric in my view, and I'll be the >first to admit that >Python is much too slow for cheminformatics. > >My preference would be for a fast, clean, back-end >cheminformatics >library >with a simple C API that could be exposed to and >interporate with >Java, C, C++, SQL, and .NET. Several proprietary >examples of this >design >exist, and they are all quite successful. >We need an open-source equivalent! >Cheers, >Warren >DeLano Scientific LLC __________________________________________________ Correo Yahoo! Espacio para todos tus mensajes, antivirus y antispam ¡gratis! Regístrate ya - http://correo.espanol.yahoo.com/
