blog: Don Marti


Open source funding and risks

03 September 2019

Read the whole thing: Recap of the funding experiment by Feross Aboukhadijeh.

Unfortunately, when open source people say things like...

Maintainers do critical work which enables companies to create billions of dollars in value, yet we capture none of that value for ourselves.

Does it have to be like this?

I’m not arguing that maintainers should start capturing all of the value that we create. But we shouldn’t capture literally none of the value either. The status quo is not tenable.

I would love to find a way to help maintainers capture at least a bit of the value we create so that we can happily continue to write new features, fix bugs, answer user questions, improve documentation, and release innovative new software.

...what people who use open source in business are hearing is more like...

We're getting a lot of software value for nothing! Fist bump!

A simple appeal to do the right thing is not something that, as a downstream user, you can put in your budget.

When you use under-funded open source software, there is always a risk that if the maintainer doesn't get paid, they will either burn out or go get a high-intensity job and let their project fall on the floor. Can you justify paying open source maintainers in order to protect yourself from this risk?

That's a little more promising, but two areas need to be addressed.

  • Is the risk quantified? I can measure a software project's value to me, but not the probability of the maintainer quitting in the absence of support, so I don't know the total size of the risk. If I can't quantify the risk, I can't justify spending to avoid it.

  • Can I measure the benefit of participating? I don't know how much my choice to fund the project reduces the risk. I could put in my $100, see that the developer still can't live on that, and end up incurring just as much cost to replace the open-source dependency as if I had not invested.

IMHO we need better market design in order to deal with those two problems. I personally think that models based on dominant assurance contracts and/or futures markets are promising (more on that later) but just banning an interesting idea after its first deployment is counterproductive.