Nevertheless, there always will be conflicts and tension between purely free open-source efforts, which are sustained entirely through volunteerism, and sustainable open-source businesses, which grow through self-reinforcing value exchange cycles (more money, more value; even more money, even more value; etc.). Yes, DeLano Scientific faces tough challenges balancing these two orthogonal perspectives, but everyone who uses any form of PyMOL should understand that the software would not be anything like what it is today were it not for the recurring contributions of PyMOL’s many sponsors over the past six years. For a direct comparison with “pure” open-source efforts, look at OpenRasMol or Jmol. They are good, but they cannot match PyMOL’s graphics.

Thus, it does not make sense to complain about DeLano Scientific doing what all businesses do in terms of marketing its products (i.e. current executable with extra featurues and documentation) and services (i.e. maintenance, support, & customization) by differentiating them from the alternatives (i.e. older or otherwise inferior PyMOL executables, free documentation, and the open-source code itself). To complain about a business acting like a business is pointless — one might as well gripe about water being wet!

Also, note that the suggestion, “There’s nothing wrong with asking people for a little goodwill, and many will happily do so especially when it’s not their own budget.” was thoroughly explored in the case of PyMOL. That exact approach was tried for two years (2003-2005) but it failed to generate enough revenue to support even a single full-time developer at a well-below-market salary.

So in the case of PyMOL at least, it is wishful thinking to believe that cost-conscious end-user scientists would voluntarily contribute appropriate value back to the project without being convinced that there is a need to do so. Indeed, the need to contribute is quite real. The assertion, “The actual value of a subscription is therefore quite low,” ignores the considerable gap which exists between raw open-source code and a ready solution to end-users’ pressing needs. Subscribers get current ready-to-run PyMOL solutions along with other benefits such as documentation. Non-subscribers get current open-source PyMOL code, older (or otherwise inferior) precompiled executables, and whatever documentation can be found on the internet. There is nothing misleading in that.

While it is true that “apt-get install pymol” can reduce the value gap on some Linux distributions, most PyMOL users are non-developers working on proprietary platforms without a compiler available. For them, the current precompiled builds available via a PyMOL Subscription and accompanied by extra documentation, features, maintenance, and support do provide *real value* above and beyond the open-source code or older precompiled executables.

Based on PyMOL’s high subscription renewal rate, such value is apparently worth paying for, and the fact that it also enables continued development of Open-Source PyMOL is a further added benefit for those sponsors who seek to advance open-source solutions while still getting their research done as quickly as possible using the latest PyMOL builds.

One can perhaps justifiably complain about PyMOL involving a tricky compromise between open-source and proprietary approaches, but it is a compromise that works and so therefore isn’t likely to change anytime soon. Thus, anyone who is put off by the PyMOL approach is encouraged to use RasMol or Jmol instead: http://openrasmol.org http://jmol.sf.net