Beancount v3 and GPL License

[Vishesh Handa]

Hello

I've noticed that one of the Goals with Beancount v3 is compatibility
with other languages, by using protobufs. This is great, however, are
there any plans to move away from a *GPL license?

I've been meaning to make a proper mobile front-end for Beancount,
which would be integrated with Git, based on the code from another
project of mine [1]. However, given that I want to ship this in iOS
where GPL licenses can be problematic [2] [3], I wouldn't want to take
the risk, even if Martin and all publicly state that it's fine.

So far, I've been reimplementing the subset of Beancount v2 format
that I use in Dart but this comes into the question of the API being
copyrightable. At some point, maybe I'll need to explicitly not
support the beancount format to not infringe on the license.

So, is there any chance we could change the license? I checked and the
"new core" part of Beancount with the protobufs only has Martin as a
contributor, so that part should be straightforward, and that way I
can easily test my parser by making it generate the same protobufs.

--
Vishesh Handa

[1] https://gitjournal.io
[2] https://opensource.stackexchange.com/questions/9500/is-apple-allowed-to-distribute-gplv3-licensed-software-through-its-ios-app-store
[3] https://www.zdnet.com/article/no-gpl-apps-for-apples-app-store/

[Matt Payne]

https://beancount.io/ advertises a mobile app in both the apple & google app stores.

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/CAOPTMKAo8t1X0FKi18ktmaKtYV8w4Xzikn48efXjOm_pYGEpvA%40mail.gmail.com.

--

"Good night, Westley. Good work. Sleep well. I'll most likely kill you in the morning." -- The Dread Pirate Roberts

[Martin Blais]

re. license, check out the history on the mailing-list for context, lots of previous discussions.

I'm really swamped - as ever, new job - and don't have time to revisit the topic for a while, let alone code on Beancount.

The problem with licensing is that it has a lot of ramifications and a lot of people have ideological commitments I'm simply not interested in.

The bottom line is that I do want Beancount to remain something other people have to share improvements they make upon up, but am not opposed to the LGPL either.

I used to know all the details intimately, but I'm a fair bit out of touch, and I honest I just don't have the cycles.

[Daniele Nicolodi]

On 02/02/2022 20:42, Vishesh Handa wrote:
> I've noticed that one of the Goals with Beancount v3 is compatibility
> with other languages, by using protobufs. This is great, however, are
> there any plans to move away from a *GPL license?

There are no plans and there would need very compelling reasons to only
start to think about something like this.

> I've been meaning to make a proper mobile front-end for Beancount,
> which would be integrated with Git, based on the code from another
> project of mine [1]. However, given that I want to ship this in iOS
> where GPL licenses can be problematic [2] [3], I wouldn't want to take
> the risk, even if Martin and all publicly state that it's fine.

IIUC Git Journal is released under the AGPLv3 license, which is just the
GPLv3 plus some extra clauses, but it is distributed on the Apple App
Store, which seems to contradict the statement that there is any problem
in shipping GPL licensed application through the Apple App Store.

> So far, I've been reimplementing the subset of Beancount v2 format
> that I use in Dart but this comes into the question of the API being
> copyrightable.

I don't see the connection between implementing a parser for a given
file format and an API. Even then, APIs are not copyrightable.

> At some point, maybe I'll need to explicitly not
> support the beancount format to not infringe on the license.

This is nonsense.

> So, is there any chance we could change the license?

I don't think so.

> I checked and the
> "new core" part of Beancount with the protobufs only has Martin as a
> contributor, so that part should be straightforward, and that way I
> can easily test my parser by making it generate the same protobufs.

I don't see how the license of the Beancount v3 parser that generates
protobufs may have anything to do with you being able to test a
different implementation of the parser against it. The license of
Beancount would matter only if you were to redistribute Beancount code,
not if use it to as a reference implementation to test your code.

Cheers,
Dan

[Daniele Nicolodi]

On 02/02/2022 21:25, Matt Payne wrote:
> https://beancount.io/ <https://beancount.io/> advertises a mobile app in

> both the apple & google app stores.

Given the very prominent but very dubious claims about encryption [1] I
would recommend to be very skeptical of all services offered by this entity.

[1] Fava operates on the ledger server side, thus it cannot be encrypted
on the server, it maybe stored in an encrypted form, but that does not
give any protection against the administrator of the server or an
attacker that gains access to the running service accessing your data.

Cheers,
Dan

[Vishesh Handa]

Hey Daniele

On Fri, Feb 4, 2022 at 2:25 AM Daniele Nicolodi <dan...@grinta.net> wrote:
>
> IIUC Git Journal is released under the AGPLv3 license, which is just the
> GPLv3 plus some extra clauses, but it is distributed on the Apple App
> Store, which seems to contradict the statement that there is any problem
> in shipping GPL licensed application through the Apple App Store.

I own the copyright of GitJournal, the parts contributed by others are
explicitly under Apache2. So, I therefore have the right to dual
license the application and ship it on the Apple App Store.

I'm planning on releasing this new app's code under GPL (2 or 3 or AGPL).

- https://github.com/vHanda/gringotts

(I'm not sure what to call the app though, any suggestions are most welcome)

>
> > So far, I've been reimplementing the subset of Beancount v2 format
> > that I use in Dart but this comes into the question of the API being
> > copyrightable.
>
> I don't see the connection between implementing a parser for a given
> file format and an API. Even then, APIs are not copyrightable.

Perfect. I wasn't too sure about APIs being copyrightable.

>
> > At some point, maybe I'll need to explicitly not
> > support the beancount format to not infringe on the license.
>
> This is nonsense.

Thank you for clarifying :)

>
> > So, is there any chance we could change the license?
>
> I don't think so.
>
> > I checked and the
> > "new core" part of Beancount with the protobufs only has Martin as a
> > contributor, so that part should be straightforward, and that way I
> > can easily test my parser by making it generate the same protobufs.
>
> I don't see how the license of the Beancount v3 parser that generates
> protobufs may have anything to do with you being able to test a
> different implementation of the parser against it. The license of
> Beancount would matter only if you were to redistribute Beancount code,
> not if use it to as a reference implementation to test your code.
>

Right, so I could copy the '.proto' files and use them as a testing
mechanism cause I wouldn't be distributing those parts?

--
Vishesh Handa