The Future of PyMOL

January 14, 2010

The pymol.org site is reporting that Schrödinger will now provide maintenance, support, and sales for PyMOL:

On January 8, 2010 Schrödinger reached an agreement with the estate of the late Dr. Warren L. DeLano to acquire PyMOL. Schrödinger will take over continued development and maintenance, as well as support and sales of PyMOL, including all current subscriptions. Schrödinger will also continue to actively support the open-source community of PyMOL.

Prior to Warren's tragic and unexpected passing, he had been working closely with Schrödinger to progressively integrate PyMOL with Schrödinger's graphical interface, Maestro. With a great sense of humility, we will work hard to pick up as best we can where Warren left off and will strive to honor his memory by continuing in the spirit and tradition of PyMOL.

There are many ways to 'sell' open source software. Delano Scientific used a freemium model; the basic PyMOL package was free, but premium services, such as support and "incentive builds", could be purchased separately. Delano took this one step further by licensing the basic package as open source.

Looking at the PyMOL license, you can see that the open source idea was taken very seriously. Although Delano Scientific didn't use a standard open source license, the terms resemble those of the very liberal BSD and MIT licenses, including the right to distribute modified copies without releasing source code (see: reciprocity). If you check out the PyMOL subversion repository, you'll notice that commit activity continues, with the most recent appearing four days ago.

The acquisition of PyMOL by Schrödinger raises many questions about the future of PyMOL. For an idea of how much interest there is in this topic, check out the Twitter PyMOL feed.

Despite these questions, or rather because of them, I think the future of PyMOL is very bright. Although there is a new commercial backer, many will have differing opinions on the best direction to take the software. The open source license under which PyMOL was developed will give these folks the power to pursue their vision. Given the widespread use of PyMOL, I would not be at all surprised to see one year from now at least one active PyMOL fork in addition to Schrödinger's. And as I've said before, forking is good.

If you've used PyMOL and/or bought premium services for it, what would you like to see in the future?

Heads-Up Credit: Molecular Modelling Blog