The sad decline of copyleft software licenses? :(

[Pen-Yuan Hsing]

(whew, finally changed my registration on this list to my FSF member 
alias address! :D)

Ugh.

I don't know if it's by chance, but recently I've read many an opinion 
claiming that copyleft free software licenses such as the GNU GPLv3 are, 
among other terrible qualities:

1. Viral
2. Cancerous (!!!)
3. Harmful to the "open source movement/principles/ethos"
4. Restrictive
5. Anti-social
6. Unfriendly
7. [other negative adjectives...]

Usually, the same people would advocate for the "permissive" set of free 
software licenses like the increasingly popular MIT license, or BSD, 
Apache, and so on. Some reasons I've seen are that these "permissive" 
licenses are easy to use so "you don't need to think about it", are 
highly compatible with other licenses, "truly open source", "truly free" 
because they don't come with any restrictions like the "restrictive" 
GPL, or that "history has clearly shown permissive licenses to be more 
successful and welcomed." Once, I asked a self-identified open source 
(not free software) advocate whether they're concerned "permissively" 
licensed code would be incorporated into proprietary software, and they 
said no: "If I create a piece of open source software, and it helps 
others even if that means it being used in proprietary products, I'm 
fine with that."

What prompted me to write this post is an academic paper I just saw:

Wilson, G., Bryan, J., Cranston, K., Kitzes, J., Nederbragt, L., & Teal, 
T. K. (2017). Good enough practices in scientific computing. PLOS 
Computational Biology, 13(6), e1005510. 
https://doi.org/10.1371/journal.pcbi.1005510

This paper is licensed CC BY (which is commendable and sadly still rare 
in peer-reviewed academic literature, though I'd prefer CC BY-SA even 
more), and IMHO it's actually very good in recommending "good enough 
practices" so that scientists who are not programmers can begin to 
improve their scientific computing practices. Quick side-note: I have an 
academic science background and I can write a thesis lamenting the 
terrible state of software development in many academic disciplines. 
Physics/mathematics/astronomy are better for sure, but scientists in 
most fields are well-meaning but horrible programmers (if they can write 
code at all, many still use proprietary GUI spreadsheets to wrestle with 
huge datasets) who don't comment their own spaghetti code that manages 
to just squeak by (usually just once, before a reboot) to produce an 
analytical result just "worthy" enough of publication. And since they 
don't really know much about programming, many genuinely think that's 
adequate. So I can say with some confidence that the paper I just cited 
is already huge progress.

To their credit, the authors of this paper emphasise the importance of 
including a LICENSE with your code, which - believe it or not - most 
other "recommendations" in the scientific literature don't mention at 
all, effectively keeping tons of scientific code unintentionally 
proprietary.

Sadly, those authors also state:

"We recommend permissive software licenses rather than the GNU General 
Public License (GPL) because it is easier to integrate permissively 
licensed software into other projects; see chapter 3 in [17]"

Where [17] is the book:
St Laurent AM. Understanding Open Source and Free Software Licensing. 
O'Reilly Media; 2004. http://www.oreilly.com/openbook/osfreesoft/book/.

The above was literally the only justification they gave for 
recommending a "permissive" license. Now, many scientists will just read 
this without critically considering its implications (somewhat ironic in 
itself) and use MIT (or similar) licenses for their code. In my opinion, 
that statement is misleading and gives the impression that the GNU GPL 
is somehow bad for code.

In case it's not already clear, I'm deeply disappointed by the almost 
complete lack of support I've seen for copyleft free software licenses. 
Other than perhaps some people on this mailing list, I have *literally* 
not interacted with any other human being who prefers copyleft licenses. 
Zero.

Also in case it's not clear, a value I hold deeply is to assume good 
faith and by default I truly respect other opinions regardless of how 
much I might disagree. Here the disagreement is regarding the nature of 
free software licenses, and it troubles me deeply.

My personal reason for desiring copyleft licenses is, in my opinion, 
simple: Software freedom is a value I hold dearly, and I want to ensure 
that freedom is perpetuated. In my view, the only "restriction" of a 
copyleft license is that it insists that freedoms are not infringed. On 
a high level, isn't a crucial mechanism for freedom to work the rule 
that "you are free to do anything as long as it doesn't infringe on 
other people's freedom?" Unless I am grossly mistaken, this is what a 
copyleft license does. Is that so bad, or "cancerous", as some might 
say? If anything, licenses such as the GNU GPL *protects* freedom, yet 
practically everyone I've ever talked to think it *limits* freedom. I 
totally understand why the FSF lists permissive licenses as free 
software-compliant, because they technically are free. But without 
protection of those freedoms, we have seen time and again how derivative 
works of free software have been locked up into proprietary products 
(with Apple's Mac OS lineage of operating systems perhaps being a famous 
example that derives much of its core from free software - some 
technically cool features, yes, but tragically proprietary and 
ultimately user-hostile).

At this point, if you also prefer copyleft licenses, I hope for your 
constructive thoughts on two sets of questions:

(1) WHAT are some common well-intentioned arguments YOU encounter for 
permissive licenses and against copyleft licenses? And what are your 
constructive responses?

(2) WHY do copyleft licenses seem to be in steep decline? What are 
positive, constructive solutions to encourage its adoption? Any big 
success stories other than the Linux kernel?

Or, if you do prefer permissive licenses, are you concerned about 
software using those licenses being used in, or made into, proprietary 
software? Are there concrete ways to secure software freedom with 
licenses such as MIT?

If there are lots of constructive responses to this post, I'd be happy 
to compile them into a FAQ of sorts and put them somewhere like the 
Libreplanet wiki or elsewhere. Ideally, I hope to foster some sort of 
curated and well-reasoned set of constructive arguments to advocate for 
copyleft licenses.

It goes without saying that I know this can be a contentious topic, but 
I know we are much better than starting flame wars and making personal 
attacks. So let's be excellent to each other. :)

Hope to learn from you!