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

Rob Landley rob at landley.net
Tue Oct 16 16:15:09 UTC 2018


On 10/15/2018 04:02 PM, Christoph Dorn wrote:> You can look at this from two
perspectives:
> 
> 1) The 0BSD name is better for those who are familiar with the BSD license
> history and want to leverage such history for meaning.

There is a well-established meaning going back decades, yes:

https://www.google.com/search?q=gpl+vs+bsd

> 2) The FPL name is better to newcomers since the name implies meaning without
> needing the history. IMO this name is nicer for newcomers.

Do you mean the part where googling for "free public license" brought up 3 other
completely unrelated licenses on the first page, or the part where:

https://www.google.com/search?q=free+software+vs+open+source

is another well-established cultural divide giving the word "free" very specific
connotations?

> I would vote for introducing a concept of "license equivalency" and alias both
> names to the same text. I have no idea about the implications of this other than
> making both names available for use.

OSI is _already_ calling it both. That's what's poisoning the google search
results for 0bsd.

> I was planning on using the FPL name extensively

Do you think it's a good _license_, or do you mean specifically the _name_
regardless of the contents of the license?

So... were you planning on writing extensive amounts of _software_ using it, or
is this some sort of marketing campaign you're just now announcing? (Do you give
talks like
https://archive.org/download/OhioLinuxfest2013/24-Rob_Landley-The_Rise_and_Fall_of_Copyleft.mp3,
have you been using the license publicly for 5 years and got it merged into
every Android system since M...?)

> as I am trying to reach a younger audience

I'm trying to do that, _and_ convince corporations to approve the use of what is
functionally equivalent to public domain code, something the OSI's own lawyer
spend years running a FUD campaign against.

> that may not be familiar with the backstory of the BSD license.
I'm not familiar with your backstory, but the website you're writing from says:

http://www.christophdorn.com/

> Realize Your Dream
>
> I am what you call an implementer of ideas and web software is my craft.
> I thrive on bringing others' ideas to life with cutting edge technology
> applied using best practices to build sustainable transactional communities.
>
> I bring solutions to the table that seem magical yet are now possible due to
> rapidly advancing web software technology.

And Google finds:

https://github.com/cadorn

> I am Binding a Software Ecosystem that uses Existing and Future Open Source
> Software as Reusable Components that are Declaratively Assembled into Systems.

In this context, I'm guessing you mean "marketing campaign".

> I don't mean to complicate this decision nor go against the well-crafted
> argument to rename to 0BSD.

So the purpose of your message was...?

> Just want to point out the two perspectives.
> 
> My reasons for using FPL over MIT (maybe I am misguided):

It's a good license, yes. What does calling it 0BSD vs Florida Power and Light
have to do with your decision?

(The other hits on Google's first three pages for "fpl" were Fantasy Premier
League, Ferguson Partners Limited, Field Programmable Logic, Forest Products
Laboratory... No mention of a software license anywhere.)

>   * It is not branded. I don't want to promote the Massachusetts Institute of
> Technology name.

Out of curiosity, what do you call "Linux"?

>   * It is shorter and to the point.

The license text is short and to the point, yes. Not as short as the John the
Ripper license:

https://openwall.info/wiki/john/licensing

Which is another BSD-derived public domain euivalent license which the author
cut down from FreeBSD. However, he removed the vast majority of the text, and my
modification only removed half a sentence:

https://github.com/landley/toybox/commit/ee86b1d8e25cb0ca9d418b33eb0dc5e7716ddc1e#diff-9879d6db96fd29134fc802214163b95a

http://lists.landley.net/pipermail/toybox-landley.net/2013-March/004589.html

And the minimal size of the delta was an explicit goal so it could easily pass
corporate approval processes that had already greenlit OpenBSD code (like
openssh and openssl, neary ubiquitously used at the time).

I described my rationale for all this at length in the github thread I keep
linking to, and yes it's something I'd been thinking about for a while:

http://landley.net/notes-2013.html#25-02-2013

I.E. I was using this license before the SPDX approval process. It was merged
into Android and shipped in Android M beta 1 before it was submitted to SPDX. It
was submitted to OSI _after_ SPDX had already approved it.

I was using this license in public 2 years before OSI ever heard about it, but I
don't get to keep my name for it? OSI's policy was to stay in sync with SPDX,
but Richard Fontana asked spdx to change its decision _after_ SPDX had approved
he license?

>   * I don't need to "intrusively" promote myself as the author by requiring a
> copyright statement.

The submission I sent to SPDX (months before Bundy) said "Your Name Here" in the
copyright statement:

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

The copyright statement is there because the OpenBSD suggested template license
had it. I.E.

https://www.openbsd.org/policy.html

Links to:

https://cvsweb.openbsd.org/src/share/misc/license.template?rev=HEAD

And that's where it comes from. And that's why the version with half a sentence
removed was called Zero Clause BSD (by analogy to CC0, which was the other
license I nearly went with when I relicensed toybox in 2011).

Alas, instead of using the template in the email they cut and pasted it off the
web URL I also sent (taking it from the live, widely deployed project), and thus
threw my name in their template. :(

Getting SPDX to fix that is on my todo list, but you see how long it's taken me
to poke _you_ guys about this...

>   * It does not require users to include a copyright notice when redistributing
> and thus eases the compliance burden (think projects composed of hundreds of
> small FPL licensed components where the project will need to include the FPL
> license only once along with a list of the components vs duplicating licenses
> with copyright statements or including authors/copyright when listing components)

Yes, the stuttering problem. I described fixing the stuttering problem as part
of my motivation in the original SPDX approval thread, _and_ explained it again
in the github thread linked above which nobody seems to have read.

Look, just search for the word "stuttering" in:

https://github.com/github/choosealicense.com/issues/464

It only occurs 3 times. (I can cut and paste it here, but I've been trying to be
quiet here because I'm not a member. I made a request, OSI board members were
debating the request, it seemed to be going well, I stayed out of it. So far the
only board member to object is the same one who tried to get SPDX to
retroactively change their decision on the 0BSD name on the spdx list in 2015.)

In that thread I _also_ explained that doing a minimal delta from an existing
BSD license (specifically the OpenBSD suggested template license, which had by
far the simplest wording and yes I'm aware it's genetically ISC but the OpenBSD
connection allows me to call it a BSD license) greatly eases corporate approval.
I linked to the Google lawyers' objections to the "public domain" portions of
musl, which unfortunately led to Google _not_ using musl in chromeos even after
Rich tracked down the original contributors and re-licensed the offending
portions of the code (although I'm told Fuchsia uses it; as an aside, I'm the
guy who convinced Rich Felker to relicense musl-libc from lgpl to BSD back in
2012, we hung out on the same IRC channels and after he read my blog entry on
https://landley.net/notes-2012.html#18-01-2012 we spoke about it at length).

Richard Fontanna was in the room when I argued with a Google developer who
insisted that CC0 was a terrible license taking away all his rights, but my 0BSD
on toybox was a much better license, and literally would not _believe_ me when I
insisted that they were equivalent. (I spoke to him right afterward.)

There's a lot of subtlety and backstory here that OSI is completely deaf to, but
sure, let's re-litigate all of it where you guys can see it.

Rob



More information about the License-review mailing list