Il y a quelques jours, Mikeal Rogers a gazouillé quelques réflexions sur la contribution à des projets libres. Je ne connais pas le contexte, mais même indépendamment de ce contexte, je pense que c’est une intéressante contribution à la réflexion pour savoir comment encourager la contribution dans nos projets
Ok, I’m starting a big thread now on the "open source funding crisis." First of all, open source won. Every developer who sits down to write a new piece of software primarily pulls from an open source stack. There are more open sources users than there ever has been. Some portion of those users could contribute in some way w/o new funding. Many open source projects lack sufficient contributors. Not due to lack of funds, but because people are choosing not to contribute. The whole "project is too big/complex" to get contributors is not true. Project status is an incentive to be involved in that project. There are also more projects people can choose to use and contribute to than ever before. Newer projects tend to have better processes. If a project is 10 years old and has a ton of users but processes from 10 years ago, contributing to that project is painful. Most developers, even in enterprise, choose what they work with/on to some degree. They are already paid and could be contributing. But the people in a position to change the policies of an older project are the ones most comfortable with the policies. Before you say "funding problem" ask that project why they aren’t on GitHub. If that’s a 2mo argument you know what the problem really is. Projects that say they are "too complex to get contributors" often have terrible docs and a filthy bug tracker, solvable w/o deep tech know. Paying full time contributors to work on projects that have policies that are keeping contributors out will likely make the situation worse. Developers who devote their time to open source aren’t lacking in equity. They are maximizing future opportunities over current compensation. However, this choice is much less accessible to huge groups of people, like people w/ families and people from underrepresented groups. We can’t fund our way to diversity by paying individuals but we can fund efforts that on-board more diverse open source users and contribs. But outreach to diverse groups only works if the policies and processes around contribution have been fixed, which you can’t just fix with $. The moment you refer to contributors as "volunteers" you’ve lost. Nobody wants to contribute for your benefit, they want to participate. This drove me insane at Mozilla. Constantly referring to "volunteers" in the "Mozilla Community." Creating any sort of separation between "paid" or "full time" contributors and everyone else will derail new contributors coming on board. This is the hardest thing for old projects: you have to optimize for the smallest contribution possible from the most casual contributor. We don’t have a funding problem. We have a culture problem. We have governance problems. All difficult to fix w/ funding. A user’s relationship to a project grows more toxic the less they can see themselves being involved, which make the project hate their users. Funding can employ people to work on community issues and, for lack of a better word, "project manage." But it’s a very hard job. Hiring people to "manage the community" without a mandate from the community is an impossible job, yet I see companies do it all the time. These strange economic models where we say "this enabled these products so it’s worth this much $" are pure insanity. You can measure the economic investment in a project by calculating the shared time investment in development but its cost is zero dollars. If you want to put it in economic terms then everyone paying the salaries of people who contribute are sharing the burden of production. This means that OSS is taking from companies the labor that is within their means to provide. "From each according to his ability" :) And that OSS is giving to each company or individual what they need at no cost. "to each according to their need." :) A future deferred payment or equity model for OSS won’t work because it doesn’t have a proper mechanism to match ability to need. Sorry to break out the socialism but this stuff is literally free, you can’t try and apply pure market economic theory to it.