[License-discuss] [License-review] Please rename "Free Public License-1.0.0" to 0BSD... again.

Tue Apr 6 09:00:54 UTC 2021

On 4/5/21 10:23 AM, McCoy Smith wrote:
> Isn't the problem here that this license was submitted initially for
> approval under "Free Public License-1.0.0"

Not "initially", no. I initially submitted it for approval to SPDX after it had
been merged into Android M, and both occured before OSI ever heard of it.

In 2011 I relicensed my toybox project from GPL 2.0 to the OpenBSD suggested
template license, calling it a BSD license when I did so:

https://github.com/landley/toybox/commit/3f998bb657bc

In 2013 I simplified it to a public domain equivalent license:

https://github.com/landley/toybox/commit/9ee76527f0fa

I then got Kirk McKusick's permission to call the result "Zero Clause BSD" when
I met him at Ohio LinuxFest in October 2013. and started publicly using that
name for it:

https://github.com/landley/toybox/commit/1748bdb6bfe4

It was merged into Android in December 2014, under that license by that name:

https://lwn.net/Articles/629362/

6 months later I was asked by a large phone manufacturer to officially submit
the license to SPDX to simplify their internal processes, which I did:

https://lists.linux-foundation.org/pipermail/spdx-legal/2015-June/001443.html

Then OSI showed up 5 months later (11 months after it had been merged into
Android and 10 months after Linux Weekly News had covered its merger into
Android), to say "I can't see OSI wanting to" and "I personally would be opposed
to", and we started arguing about it:

https://lists.linux-foundation.org/pipermail/spdx-legal/2015-November/001568.html

(That was the first objection I ever encountered that this wasn't really a BSD
license: after 4 years of publicly using it, giving a talk about the project at
the embedded linux conference in 2013 that described the licensing at length
https://www.youtube.com/watch?v=SGmtP5Lg_t0 and a dedicated licensing talk at
the conference where I met McKusick
https://ia803008.us.archive.org/34/items/OhioLinuxfest2013/24-Rob_Landley-The_Rise_and_Fall_of_Copyleft.mp3
and nobody ever said "I object that the license you're using is not a BSD"
before OSI looking for a reason not to change their existing decision. The
reason I asked McKusick for permission to call it "zero clause BSD" was because
I wasn't part of the BSD community and was creating a new member of the family
alongside "4/3/2", not because the license I was using hadn't come to me through
OpenBSD's website. I wanted it blessed by an old hand, and it was, in 2013.)

I posted a timeline like this as part of the 2015 OSI-on-SPDX argument:

https://lists.linux-foundation.org/pipermail/spdx-legal/2015-December/001574.html

And I posted that last link in the 2018 OSI naming decision thread when I came
here to ask OSI to eliminate the market confusion:

http://lists.opensource.org/pipermail/license-review_lists.opensource.org/2018-October/003609.html

> therefore in order to understand
> the debate about the approval,

Why would they need to? OSI got caught in a
https://www.scientificamerican.com/article/the-brontosaurus-is-back1/ style
issue and at the time had no procedure to change an existing decision. That's
the main reason (they stated at the time) that fixing it was a big deal: they
had to come up with a new procedure.

> one would at least need to reference that
> name so that anyone interested in the discussion about approval would know
> to look under that name?

You don't have to know any name other than 0BSD and that it was approved by SPDX
and OSI to find the actual discussions in the web archives, and I know this
because I just Googled the above links.

There's a bunch of great history in those approval threads about related issues
such as why CC0's initial submission wasn't approved by OSI:

http://lists.opensource.org/pipermail/license-review_lists.opensource.org/2018-October/003627.html

But that information does not belong on the 0BSD summary page either.

I myself have given summaries of this information in plenty of other places,
such as the github approval process for adding 0BSD to choose-a-license, mostly
in two chunks, as a conversation with Christian Bundy who submitted it under the
other name to OSI:

https://github.com/github/choosealicense.com/issues/464#issuecomment-289236817
https://github.com/github/choosealicense.com/issues/464#issuecomment-289526758

The reason that github approval process failed at the time was the "Market
confusion" over the two names, which is why I came here to OSI to resolve that
in 2018, which resulted in 0BSD successfully being added to github:

https://github.com/github/choosealicense.com/pull/643

Which happened precisely because OSI had changed its page, allowing wikipedia to
change, which stopped causing confusion that prevented license adoption.

There are now twenty five thousand repositories on github using 0BSD:

https://github.com/search?q=license%3A0bsd&type=Repositories&ref=advsearch&l=&l=

And that search is _not_ including forks.

> I'm not sure this is a problem that OSI created.

I am sure. I was there. More to the point, it is a problem OSI decided to fix,
but it's not currently fixed. Here is the 2018 decision:

http://lists.opensource.org/pipermail/license-review_lists.opensource.org/2018-November/003830.html

> The OSI board has decided to rename the Free Public License 1.0.0 to
> "Zero-Clause BSD". The opensource.org website will be updated in due
> course.
> 
> We did this to address concerns about potential confusion over and
> inconsistency surrounding the use of multiple names for the same
> license text.
> 
> The original submitter of the FPL, Christian Bundy, indicated in
> private communication that he did not object to the renaming.

OSI announced a decision to fix the problem and a rationale for that decision.
Not "dual name": removing second name because multiple names bad. Unfortunately
the first fix OSI implemented was incomplete, and over time the band-aid came off.

You're now arguing OSI can't fix the problem because then people wouldn't know
there had been a problem.

> Also: it looks like this license was being called Free Public License 1.0.0
> elsewhere several years ago, so I'm not sure it's OSI that did the renaming
> (or improper naming).
> https://tldrlegal.com/license/free-public-license-1.0.0#summary

No, they got that from OSI. That kind of feedback loop propagating
misinformation is called "citogenesis":

https://xkcd.com/978/

It's also what happened to wikipedia earlier this year, and even seems to have
been what occured WITHIN the OSI website, where someone saw the historical note
and moved the information into the title. That is why it needs to be removed
entirely (consistent with the 2018 decision, and the removal of the old URL to
the license summary page under the old name), so it doesn't happen again in a
couple years.

Part of my research back in 2015 was attempting to confirm that no software had
ever shipped under the "Free Public License". It's hard to prove a negative, but
I couldn't find any: it was proposed for use in new software, but had not BEEN
used by any software. Meanwhile Android M including toybox shipped October 5,
2015 (a month before an OSI representative showed up on the SPDX list), with an
annual sales volume of one billion units. (That doesn't make it "better", but at
least in theory makes it easier to notice its existence.)

I repeated that search during the 2018 thread:

http://lists.opensource.org/pipermail/license-review_lists.opensource.org/2018-October/003609.html

This issue has been re-litigated in 2015, 2018, and you're making the same
objections in 2021. Can OSI please fully retire the "Free Public 1.0.0" name so
we don't have to do this again in 2024?

Thank you for your time, sorry the reply is so long.

Rob