[MMTK] Re: [PyMOL] improving interoperability between python moleculetoolkits (fwd)

Eugen Leitl eugen@leitl.org
Thu, 12 Dec 2002 18:16:38 +0100 (CET)

-- Eugen* Leitl <a href="http://leitl.org">leitl</a>
ICBMTO: N48 04'14.8'' E11 36'41.2'' http://eugen.leitl.org
83E5CA02: EDE4 7193 0833 A96B 07A7  1A88 AA58 0E89 83E5 CA02
http://moleculardevices.org         http://nanomachines.net

---------- Forwarded message ----------
Date: Thu, 12 Dec 2002 12:02:20 -0500
From: Richard Gillilan <reg8@cornell.edu>
To: Jules Jacobsen <jobj2@hermes.cam.ac.uk>
Cc: Michael Sorich <mike_sorich@hotmail.com>, pymol-users@lists.sourceforge.net
Subject: Re: [PyMOL] improving interoperability between python

Jules Jacobsen wrote:
... stuff deleted
> The core of this program is the data model in which each atom in a
> molecule is defined. The next layer are the conversion scripts which take
> the molecule data from the model and pipe them out to whatever format is
> needed for the various data manipulation programs (Aqua, XPLOR and the
> like). Seeing as this model is in existance and will be tightly integrated
> with the PDB databank for their structural checking procedures prior to
> accepting a model it would make sense that this be used as an industrial
> standard.

My two cents:

One of the reasons I have stuck with OpenDX (www.opendx.org) for visualization for so 
long is the general (object-oriented) data model. It has its limitations,
but it has proven very powerful. Users can define a wide variety of data in 
hierarchical forms. Modules can be quite smart about the data structures they
recieve: for example the isosurface module can accept groups of 2D and 3D grids
and will produce corresponding groups of lines and surfaces. It was also easy
for me to create a RenderMan/BlueMoon module that allowed OpenDX users to seamlessly
access almost all the features of the renderer without modifying any OpenDX code.
Same was true of the MatLab module a student and I created.
This was all possible because of the general data model (which, by the
way, was designed with parallelism in mind from the start).

I'm not neccessarily trying to plug the OpenDX model (though I strongly
enourage folks to take careful look), but I think that any future system 
should put a great deal of work into developing a strong underlying
general data model, not just for molecules, but for all kinds of scientific data:
grids, vector fields, polygonal surfaces, trees etc. Just leaving things open to 
general objects is not a good idea ... there needs to be some structure and standardization
so that 3rd-party modules can be intelligent and interoperable.

Anyone ever checked out the Object Management Group (OMG)? I went to one of their
meetings (Objects ion Bio- & Chem-Informatics 2001) a while back and was impressed
with the concept. They are basically a non-profit consortium that sets open industry 
standards for objects across all different languages. CORBA is one of their babies. 
The PDB is also involved with OMG as are a number of bio/macromolecule groups (Michel
Sanner of Scripps gave a nice talk on his Python-based modeling system). Worth looking 

Richard Gillilan
MacCHESS, Cornell

This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility 
Learn to use your power at OSDN's High Performance Computing Channel
PyMOL-users mailing list