Nat & others, Thanks for your input.
>and I would be happy to contribute some of my time. However, my >impression has been that future PyMOL development will be almost entirely >centralized. Not at all. Here is an overview of DeLano Scientific's Open-Source vision: Our plan is to eventually distribute development and support as broadly as possible, throughout the world. However, PyMOL's current C core is such spagetti code that a distributed approach would surely fail with that module. Other developers who have looked over the source code have shared this opinion, and there have been very few core code contributions in past years. Thus, for the near term, I expect to continue maintaining and improving the core myself with help from a few people who can work closely with me (in person, with a white-board, correcting my mistakes, preparing high-quality documentation and robust specifications). Nevertheless, much distributed development can and will happen through our *main* strategy in which remote developers assume responsibility for well-defined, standalone modules that are wrapped behind simple Python interfaces. The great thing about doing development in this way is that those modules will also find utility and adoption beyond PyMOL. Standalone Python (and C/C++) modules will eventually provide PyMOL with useful capabilities such as: - Custom GUIs, built using Tkinter, wxWindows, QT, Jython, or ActiveX. - an electrostatics engine - diverse file I/O - batch movie generation - one or more molecular mechanics force-fields - a gradient minimization engine - homology model generation - a protein morphing engine - a sequence alignment engine - a SMARTS/SMILES-enabled chemical database engine - a secondary structure assignment engine - crystallographic target functions - 2D->3D conformer generation - guided tutorials and so on... In each case, these modules will communicate data to and from PyMOL via Python interfaces. Opaque C or C++ objects will be used to pass large data sets that Python couldn't otherwise handle efficiently. Yes, core development must remain centralized (at least for now), but we are absolutely planning to decentralize development of most of the new and exciting features which will come into existence in the next few years. My hope is that we will eventually raise sufficient recurring sponsorship funds to compensate a whole fleet of remote, part-time developers for their work. Would developer-scientists be interested in earning some extra income writing Open-Source software according to published guidelines and specifications? I hope so, and particularly so when they are able to retain ownership of the code, giving us just an unrestricted license to use, modify, and redistribute it, in the same way we make PyMOL available to all of you. Eventually, we hope to compensate numerous parties engaged in Open-Source development. Likewise, the plan is to spread many of the support responsibilities over a base of knowlegable users. The details are sketchy, but I know for certain that the best people to support the program are the same people who are using it to solve real-world problems. As with developers, the goal is to provide reasonable compensation to those who participate in this capacity. In the meantime, the volutary/collaborative approach seems to be working reasonably well, and I continue to thank active members of the mailing list for their wonderful contributions. Please understand that the above outline is just a general picture of how things are likely to shape up, and that the details are subject to change. It may take a decade to reach this scale of operation, but I am confident that DeLano Scientific will eventually get there, once people come to appreciate the research-enabling potential of Open-Source. All things in time. Assembling a decent manual is top priority right now, and then we need to make sure that we are providing sufficient incentives/rewards to motivate continuing sponsorship of the project. We will need substantial backing from individuals, governments, academia, biotech, and the pharmaceutical industry in order to realize this Open-Source vision. As always, your thoughts are welcome. Cheers, Warren -- mailto:war...@delanoscientific.com Warren L. DeLano, Ph.D. Principal DeLano Scientific LLC Voice (650)-346-1154 Fax (650)-593-4020
