Public Goods Are For The Whole Public

This page discusses the particular issue of programmer-centric and business-centric elements of the Open Source world via a critical review of an outside report on public software infrastructure.

Programmer and business centrism in Open Source software

Of everyone who understands and directly deals with FLO software, a large portion lives within the insular worlds of programmers and tech startups. From their perspectives, Open Source code is for coders. It’s about making life easier for businesses who use Open Source infrastructure and about easing the path for new programmers. All too often, people with this perspective ignore the issues of power and economics in how the software ends up affecting the rest of society who receives the technology downstream.

A quintessential example of this narrow view comes from Nadia Eghbal who received funding from the Ford Foundation to write her lengthy but slick, professionally laid-out report about the issues of funding FLO software as public goods: Roads and Bridges: The Unseen Labor Behind Our Digital Infrastructure

Like (which she was aware of but does not mention in the report), she embraces the metaphor of roads and road maintenance to talk about public goods and the challenges in funding them. Many of her points are the same ones we’ve been making throughout our publications and research in the development of The primary difference is the extent to which Nadia talks only about upstream software and ignores the question of whether the very same points apply to the downstream software made by the same companies she references as users of the infrastructure. articles about public goods and software funding

For reference, the particularly relevant posts we’ve made:

Our review of Nadia’s report

At, we care deeply about these issues and want to see them more discussed. We want everyone to embrace this metaphor of public goods (particularly public roads) that we made the cornerstone and the very name of when we started the project back in 2012. Our name refers to the Snowdrift Dilemma, a specific road-based game in the study of public goods (essentially: who will clear the snowdrift from the road we all want to use?)

Effectively, we agree with Nadia that FLOSS is a public good that needs public support. Of course, roads are typically funded by taxation. Some of us would be happy to see (and pay) a tax to fund FLOSS, but we’re not in a position to lobby for that (not to mention anarchist and libertarian arguments against the concept of mandatory taxation in general).’s crowdmatching system can be considered a substitute for taxation — the best we can manage with a voluntary mutual-assurance contract to fund public goods. So, we’re addressing the very issue of public infrastructure that Nadia talks about.

The main issue with Nadia’s perspective is that it simply isn’t going far enough. Her report is written as though public roads and bridges are primarily of interest to commercial businesses. Of course, all citizens use public roads which are just as important to individuals and local communities as to business ventures.

Applying this metaphor fully to today’s software landscape, we see that the vast majority of all the roads (programs) that regular citizens drive on (run on their devices) are proprietary. The public options pale in comparison with only a few exceptions. To describe the problem only in terms of the business interests would be comparable to saying that we need only repair the roads and bridges that lead to industrial sites; that we should be fine with privatizing all neighborhood streets and non-business-centric infrastructure. Essentially, that’s what we have today: privatized proprietary software nearly everywhere and inadequate funding for both the back-end FLO software that businesses use and those downstream end-user programs which are FLO public goods.

So, here’s a clear visual of the options we have today:

Proprietary toll-roads with ads and surveillance versus inadequately maintained public roads

Proprietary toll-roads with ads and surveillance versus inadequately maintained public roads

We can use restricted, crappy toll-roads surrounded by billboards and surveillance cameras (i.e. the ad-based invasive proprietary programs and web services) or we can struggle with inadequately-maintained, unreliable public roads (the half-done poorly-funded FLO software). We absolutely need to fix this and create a sustainable approach for funding public roads (software), and we should all wish to eliminate the counter-productive restrictions of the privatized roads/software. Our efforts need to extend to all software, not just the software that serves commercial interests.

Nadia writes:

“Because code is less charismatic than a hit YouTube video or Kickstarter campaign, there is little public awareness of and appreciation for this work.”

That description is far removed from the world we live in. We see people as enthusiastic about their smartphone apps and favorite websites and other end-user software as they are about YouTube videos. The distinction Nadia glossed over is whether we’re talking upstream or downstream. OpenSSL isn’t comparable to a YouTube video; OpenSSL is to a popular website what a camera lens is to a YouTube video. It’s true that a website visitor doesn’t care about OpenSSL, but neither does a video viewer care whether the video maker’s camera was built with sweatshop labor.

The exploitation of backend Open Source software by proprietary interests isn’t all that different from the exploitation of any core resource by capital. We would do well to discuss the similarities (and differences) between complaints of exploited FLO software developers and those of exploited physical laborers around the socialist and union movements in the 20th century (a complex issue too much to say more here). Achieving any real sustainability requires at least acknowledging the core nature of such exploitation.

Nadia makes many wonderful and valid points, and they just apply further than her limited framing. She writes:

“If such critical pieces of software were not free, people from all walks of life would not be able to take part in today’s technology renaissance.”

Yes! But we could just as well remove “technology renaissance” and replace it with almost any other field: science, music, art, history, public policy, community organizing… Everything today uses software, and the same great opportunities that FLO software provides to budding developers could be available to students of all fields, if only the software for those fields were FLO public goods too! Of course, they’d face the same challenges in funding and all…

Nadia even hints at this with sentences like “Although the “serfdom” argument mostly refers to a company’s products, such as an iPhone…” — but those things are software infrastructure too! They just are infrastructure for people who do things other than programming. All the same concerns apply. We need to get beyond the framing that appears to pretend that FLO public goods issues don’t apply to downstream software. It’s indeed interesting how the “post-Open-Source” generation of software developers have myopic views seeing the whole FLO issue only from their developer perspective, but that doesn’t mean we should accept that any more than we accept myopic chauvinism from people who assume programming is for white men. It’s a problem that we aren’t doing enough to show today’s GitHub contributors how much the same FLO, public values apply to everyone else including the end-users.

Side note: Nadia says, “[GitHub] being proprietary is a controversial topic for some open source contributors.” Well, it would make sense to at least reference a footnote to explain why people see proprietary platforms for free software as a problem, but aside from those philosophical views, the more important issue here was glossed over: that GitHub being proprietary is the most obvious point of contrast showing the issue of FLO versus proprietary overall. It’s because of this starkness that the topic comes up, but the issue is the same if you apply it to any web service (i.e. that proprietary terms deny us all the value that FLO public projects have but have a more reliable funding model)

Side note 2: From Nadia’s report: “…more restrictive AGPL license that would require attribution…” What‽ That is not at all a good description of the AGPL. Attribution is not the issue here. The entire issue is that AGPL requires keeping the FLO license terms for any public use of the software including for derivatives (see our licensing discussion). In a case of paying for MIT instead of AGPL, businesses aren’t paying for the convenience of avoiding attribution requirements. They’re paying for the power to opt their product out of the public goods infrastructure. Describing this any other way seems like an effort to keep readers from ever thinking about copyleft. The potential value of copyleft in sustaining the whole ecosystem is a major topic that Nadia’s whole paper left out. Copyleft specifically blocks the sort of exploitation in question without violating the Free Software or Open Source definitions. To only reference it in context along with non-FLO shareware seems disingenuous.

Regarding the forward-looking points and suggestions, they are basically spot on. Nadia has done a great job elucidating these. Again, we just need to apply them to all software, and really all non-rivalrous soft things (music, research, art, etc).