In the conversation, we got to applying this test. Concrete examples will force us to clarify the test itself so we end up with reliably precise questions. Examples off the top of my head:
- Non-rivalrous to read (ignoring near-zero bandwidth server costs), but there’s rivalry around whose edits get accepted and stable officially (we can’t both edit the same sentence in conflicting ways and get our version on the main page). Take away: we need to clarify that the rivalrous question in general is fuzzy and doesn’t apply to the issue of publishing a specific recognized/endorsed/official edition.
- Use with freedom? check. Study? check (we can see the raw markup, the history, and also the source code of the software that runs it) Modify? check (besides modifying directly, we can modify our own copy, and that is the test we care about). Share? check (it’s fine that it has the Share-Alike restriction that says you can’t remove the freedoms when you share it, but otherwise, sharing is unemcumbered and non-exclusive, anyone can copy and share.
So, updated fleshed out test questions:
Is it rivalrous or exclusive:
- to use?
- to study?
- to modify a copy?
- to share (i.e. publish and copy for others, both original or modified versions)?
Let’s consider another example:
- Public park?
- Use? There are often some restrictions on when use is allowed and restrictions on what the park may be used for, so it’s fuzzy and not unambiguously free from usage restrictions. It’s rivalrous at a point where it gets crowded or one type of use (e.g. some particular game) limits use for others
- Study? sure
- Modify (a copy)? Sure, if you copy a park by making another with the same design, you are free to modify that copy, but copying has practical feasibility issues. Modifying the original park itself is rivalrous to whatever extend it’s allowed anyway (which is usually limited)
- Share? Yes
Okay, parks are clearly fuzzy and have a lot of aspects of public goods but have rivalrous aspects which make them more like common-pool resources, rivalrous but non-exclusive.
- Public radio (e.g. NPR in the U.S.)?
- Use? Sure, everyone can listen at once, no rivalry, no exclusions
- Study? Well, how far does this freedom go? Do we need to have access to the unedited source materials to study how the stories were put together? Studying is free for the primary published material, and unlike software programs, we don’t really need the source material to know what the story is doing or how it works.
- Modify (a copy)? Sure, privately copying and editing radio stories isn’t blocked for anyone
- Share? NO! Even public radio has All Rights Reserved terms that require special permission for anyone to publish the stories. Each affiliate station pays for the right to do this. And sharing modified versions is only allowed up to where it counts as “fair use” under U.S. law.
So, public radio is a club good even though use is non-exclusive. The freedom to share is limited to an exclusive club.
Perhaps the Open Content 5 R’s could be a good alternative test. Retain, Revise, Remix, Reuse, Redistribute. Another good formal definition is the Open Definition which defines “Open Works” Importantly, the Open Definition highlights access which all the other definitions seem to ignore or sideline, but that is key to public goods. It’s not enough to have freedom to use if accessing something at all is restricted or artificially limited.
We already reference the Open Definition among other definitions for required terms for projects at Snowdrift Wiki - Project Requirements
The question here is how to boil down these formal terms into something that is accessible to the general public and successfully pumps their intuitions.
I suggested that there are unambiguous public goods and semi-public-goods (or partially-public, however we want to say it).
So, we’re looking for a question everyone can easily ask even if they aren’t comfortable answering it at first. Something like “can anyone and everyone freely Retain, Revise, Remix, Reuse, and Redistribute the work?” If so, that counts as public goods.
One concern: I want people to focus on the economic nature of public goods, not just the question of whether publications are FLO. The FLO-digital-works test questions might be too narrow. Cleaning up the Pacific Garbage Gyre is economically a public goods issue. The result of the work benefits the whole world not some exclusive group, the value is non-rivalrous and non-exclusive. But it doesn’t fit the 5 R’s or 4 freedoms. The important thing is that it doesn’t fail the test questions. It just doesn’t apply. Use of a healthier world isn’t rivalrous or exclusive, nor is studying, modifying copies, or sharing… it’s just that copying isn’t something that applies at all.
So, maybe the test should focus on whether there are restrictions.
A carefully complete test
I’ve updated the test with my preferred and precise language for all the important points:
Is it rivalrous? I.e. does one person enjoying the work reduce its availability or value to others?
Are any freedoms reserved to an exclusive club? Particularly the freedoms to access the resource, to retain copies, to revise copies (with access to enough source material that revision is practical and not hampered), to remix, to use and reuse, and to share (both original and modified versions)?
I don’t like the verboseness of the explanations, but the two key questions are pretty short and clear. I didn’t include the freedom to modify the original/official version as that seems inevitably rivalrous, so while that could be either exclusive or non-exclusive, it can’t be a public-goods feature (or else nothing would be public goods).
Wikipedia passes the test.
Parks can pass the test only with notable qualifications (rivalry depends if there’s a risk of crowding; there’s no restrictions on the freedoms but they mostly are non-applicable)
Public radio fails the test (freedom to share is exclusive)
Environmental clean-up passes the test (with the caveat that the freedoms are non-applicable)
Let’s add one more: GitLab partially passes the test because the FLO part of their work passes, but the proprietary features are exclusive, so as a whole project, it fails the test.
We aren’t planning to fund stuff like parks let alone environmental clean-up work. But we could in principle expand to those things because they do have the core economic public-goods situation that we’re focusing on addressing. Clearing a snowdrift from the road is itself in this category. So, I think explicitly including this in the scope of the intuitions we give newcomers is important. It’s not enough that people understand that we’re funding FLO digital works, our priority is that people notice that FLO digital works are a type of public goods, and they need a concept of public goods in order to see that and then really have it click what the issues are here.