Using (Debian) Money to Fund Debian Projects ============================================ Related documentation: “Problems and Strategies in Financing Voluntary Free Software Projects” by Benjamin Mako Hill https://mako.cc/writing/funding_volunteers/funding_volunteers.html Known problems when funding the work of persons ----------------------------------------------- * Jealousy between members of the (development) team, when one is paid to work on something when the others is not. * Tendency for volunteers to stop doing something when a paid contributor is doing the same kind of work (either because the paid contributor is more responsive or because the volunteer uses this as an opportunity to change their own priorities). * Paid contributors who have been given a goal and no clear guidelines/rules will pick the “easy way“ and will (unconsciously) make it harder for other people to help with the work or to take it over afterwards, etc. * Paid contributors might work towards the specific needs of the sponsors at the expense of the generic needs of users. Generic recommendations to limit the problems --------------------------------------------- * Fund work that volunteers do not do (and that volunteers recognize as work that no volunteers would want to do). Administrative work (project management, book-keeping, etc.) tend to fall in this category. - Debian LTS is in that category. * Fund limited projects, in scope and duration. - Google Summer Of Code are in that category. - Debian LTS is entirely the opposite, it is funding recurring security work. * Fund features that have been waiting in TODO lists for a long time and that are widely recognized as being important. - Debian PPA would be in that category. * The process to decide who/what gets funded must be fair and transparent (as much as possible). - Debian LTS has multiple rules (public reports, clear criteria to join the set of paid contributors). * Fund work through external organizations. - Debian LTS also made that choice with Freexian (a private company) collecting sponsorship and dispatching work hours (and not SPI or any trusted organization). Other Free Software Project Paying Developers --------------------------------------------- * Django Foundation paying Django Fellows https://www.djangoproject.com/fundraising/#fellowship-program * Tasks labeled as “administrative and community management tasks” - handling security@djangoproject.com - fixing release blockers, helping to ensure timely releases - fixing severe bugs (and backporting them) - reviewing and merging pull requests - triaging tickets - answering user questions on IRC and mailing list - helping new Django contributors to land patches and learn the project's philosophy (from https://www.djangoproject.com/weblog/2015/feb/06/call-applicants-django-fellowship/) * Weekly report to few members of the Django core team as well as public reports on the mailing list: https://groups.google.com/forum/#!topic/django-developers/bkBOsFJqHMQ * Yearly retrospectives: https://www.djangoproject.com/weblog/2015/jan/21/django-fellowship-retrospective/ https://www.djangoproject.com/weblog/2015/dec/17/fellowship-2015-retrospective/ => very positive results * Reproducible builds * Jérémy Bobbio (Lunar) and Holger Levsen paid by Core Infrastructure Initiative: https://www.coreinfrastructure.org/grants * Tasks: "steering the reproducible build effort" ... * Tails * Few public details but their accounting shows significant amount of money (> 50K USD) used in "Subcontracting": https://tails.boum.org/doc/about/finances/index.en.html * They fund boring administrative tasks, and this includes fundraising itself. * When they fund development projects, they are related to their roadmap which is elaborated by the community. * No formalized process. * Perl Foundation * Funds "random" development, and specially maintainance, projects for limited periods * http://www.perlfoundation.org/how_to_write_a_proposal * FreeBSD has a number of developers who privately searched funding for particular projects. In some cases, the FreeBSD foundation has provided funding for those projects once a critical mass of funding had been provided through other means. First example of this at http://people.freebsd.org/~phk/funding.html Questions to the audience ------------------------- gobby infinote://gobby.debian.org/debconf16/bof/funding-debian-projects 1/ Besides Debian LTS security work, are there other tasks that we should do in Debian and that we are not doing because not enough persons are motivated by them? - fundraising (currently mainly done by DebConf folks), needs to be coordinated across all entities involved. There are many aspects of fundraising that can be outsourced, but we ought to be very careful when incentivising people to raise funds. - trusted organisations are mainly treasuries for the Debian project under control of the DPL. They may have their own accounting standards to meet but these are likely not going to be helpful for Debian to communicate to sponsors how we use their funds. - accounting consolidated over the project allows us to make meaningful state- ments about the use and effect of funds. Moreover, involving a contractor brings benefit of introducing an interface (discipline for us, accountability for them) with the external (numbers) world - auditing then is the final step, - any mentoring/outreach effort where we send Debian contributors to recruit new persons - designing themes and artwork (with prizes?) 2/ Do you know of widely desired features that we have been waiting on for a long time? - PPA/bikesheds (why bikesheds in this context? it's not a "feature" we want!) - "bug cleanup" on important packages - https://wiki.debian.org/BudgetIdeas - wiki consolidation - chasing down people for relicencing 3/ If Debian had its own "grant program", similar to Google Summer of Code, to fund scope- and time-limited projects proposed by Debian contributors, how should the selection process look like to match the fairness and transparency requirements that are needed? - the Perl Foundation grant program might be a good model for this - community review before any project goes through grant committee - grant committee delegated by the DPL, restrictions - try it as an experiment with clear questions that we want to answer - might be easier to use money out of cash flow rather than the Debian savings but it'd be recommended to bootstrap the programme, rather than expect them to raise their own funds before they even start. 4/ If there was such a program, with applications open to all Debian contributors (so no restriction to students-only), would some of you propose projects that you would implement if you were funded to complete them?