On Mar 27, 2019, at 08:24, Francois Berenger <mli...@ligand.eu> wrote: > As an open-source project, I feel rdkit is quite successful. > So, the user community is not so small. > Some people who cannot contribute time could contribute money to the project > (especially if it is tax-deductible, I guess).
I think the questions are "why would they contribute money?" and "why haven't they contributed money?". If those questions cannot be answered well, then there's little reason to go further down this path to the next question, which is "how do we effectively encourage them to contribute money in the future?". To be clear, Novartis contributed a lot of money for the RDKit development. Roche also funded me to develop and contribute the MCS package now part of the RDKit core, and the mmpdb project which was contributed to RDKit. These are also financial contributions and must not be ignored, and these are not the only two organizations which have done that. But I honestly thought that there would be more interest in hiring my services as a consultant, to work on further development of open source software. I feel like there are clear economic benefits for companies to fund open source packages. Instead, it feels like the more open source software packages I write and release, the fewer leads I get for new consulting work, compared to when I gave "I wrote this in-house application for company X that no one else will ever use" talks. Perhaps what's easily available for no cost is seen as having no value, while that which is hidden, no matter how hacky, is treasured? My optimism started 20 years ago, when I was still involved with the Biopython project. My company offered commercial support for Biopython, and I had NDAs in place with several of the other Biopython developers so we could easily be funded to work on specific improvements that an organization might need. I never found someone interested in providing that sort of funding for Biopython, and it still looks like that's the case in cheminformatics. See also 'Roads and Bridges: The Unseen Labor Behind Our Digital Infrastructure' (ref. 53 in my paper) for further examples of the difficulties in funding open source work. https://www.fordfoundation.org/about/library/reports-and-studies/roads-and-bridges-the-unseen-labor-behind-our-digital-infrastructure/ > On Mar 27, 2019, at 10:06, Greg Landrum <greg.land...@gmail.com> wrote: > If rdkit was accepted at the software freedom conservancy, I understand > the management fee would be 10%: There's also Software in the Public Interest, which "serves the free software and open source community by facilitating the administrative and financial needs of its associated projects", including the Open Bioinformatics (ex-)Foundation. When the OBF was created, it was common for many groups to start their own foundations. Since most of the administrative needs are the same for the different projects, it makes sense to consolidate. > A question since I genuinely don't know: is it important to anyone that this > go through a not-for-profit entity? The OBF became a not-for-profit to make it easier to organize the BOSC (Bioinformatics Open Source Conference) meetings. Some of the early BOSC meetings were run out of someone's personal bank account, and he was personally financially liable in case of problems. Working through a non-profit makes it easier to set up things like summer internships (a la Google Summer of Code) and travel support, because the payment is less likely to be viewed as a way to get around employment laws. Open Bioinformatics has a Travel Fellowship program. I don't know the details. Looking at the report for 2018 at http://spi-inc.org/corporate/annual-reports/2018.pdf , Open Bioinformatics spends about $5,000/year for IT and meet ups, an "ordinary income" of $5,400, and an equity of $85K. There's overhead to running a non-profit, like filing paperwork, and that requires specialized knowledge. For revenues that small, it really helps to be affiliated with an existing umbrella organization. The OBF gave up their incorporation in 2012 to be an SPI-associated project. For what RDKit does now, I see no need to set up/join a foundation. T5 Informatics can organize an RDKit UGM the same way that any vendor can organize a UGM, and company acts as the firewall to your personal finances. T5 (or Dalke Scientific :) can also act as an intermediary if, for some reason, a company does not wish to fund someone directly. Though you'll have your own overhead in that case, because of the additional tax requirements in dealing with subcontractors. No matter what, that's going to be easier than arranging things through a university, even Paul's. It's only if you start getting multiple organizations interested in contributing funding, or really want the transparency that T5 and RDKit funding are not intermingled, where I would suggest looking at that. Another reason is if you want RDKit-the-organization to exist develop on its own, even without you. For example, Guido van Rossum is not the head of the Python Software Foundation nor on its board. Cheers, Andrew da...@dalkescientific.com _______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
