Contributors mailing list archives

For me AGPL is the best license. IMHO it should be the license choice for everything including Odoo CE.

For clients it is the best (despite all the FUD in this email chain). The simple reason is clause 9, no requirement to accept as they are not propagating or modifying. This renders clause 13 moot which is what everyone gets their knickers in a twist about. Way back when in 2015 we talked to a lot of people including the author of AGPL, engaged specialist lawyers etc this was the basis that you see in our current policy. Furthermore, insofar as a partner/developer develops dependent modules, they are now bound to provide their client the source code, even though client is not required to accept licence, developer is. So client is even further protected and avoids lockin.

For (at least open source minded) partners/developers it is the best. You can develop extensions, do modifications and convey them to your client, along with source and yes it is licensed AGPL. But if someone extends your work and puts on App store you have some protections. It allows the software to grow and evolve and encourage contribution. Whats more is you can extend LGPL without fear the LGPL author will take your contributions. OpenOffice vs LibreOffice success is testament to that.

For Odoo it would be best. It would protect them from SaaS startups under clause 13. Now I hear you say, but enterprise is incompatible so it must be LGPL. This is simply not true, insofar you believe that Odoo correctly relicensed and owns copyrights to all the source. All they needed to do was dual license.

So why do we end up with emails like this, Odoo's decision etc. Because SaaS providers, and vendors like Microsoft have done such a good job of spreading FUD, a lot of which is repeated here that everyone is afraid of it, when it should be embraced. Literally the only people that should fear AGPL is unscrupulous SaaS providers and public for sale app developers. 

Insofar as clients need protections around "proprietary code" being given to other people, that is a contractual issue between them and their partner. But let's say your client has really bought the FUD, LGPL is still a bad choice because it forces client acceptance, GPL would be better which is basically the AGPL without network clause.

On Sat, Sep 13, 2025 at 9:22 AM Enric Tobella Alomar <notifications@odoo-community.org> wrote:

Sure,

If I remember properly, we have 3 repositories that are mainly LGPL: DMS, EDI-Framework and Queue Job

I will take EDI-Framework as a base as it has several modules in the repo. It has 17 modules in 18 and it as been in its repo for 5 versions for a total of 197 PRs. this makes an average of 39 PRs for each version. Also, it has 49 contributors

On the other hand I will use helpdesk as it has a similar number of modules in 18 (17) and it is completly AGPL. It has been there for 9 versions for a total of 686 PRs, that makes an average of 76 PRs for version with a total of 120 contributors

It is true that EDI-Framework has some special cases like the complexity of components and so on, but with similar sizes in number of modules we can see quite a difference in number of contributors and PRs done.

I know that correlation doesn’t imply causation (spurious relationships are one of the first fun lessons in statistics), but in my view, this makes it quite clear that licensing alone is not the decisive factor in how contributors engage with a project.

Kind regards

El vie, 12 sept 2025 a las 21:57, Frederik Kramer (<notifications@odoo-community.org>) escribió:

Hi Enric,

ok, i obviously didn't recall this fact. In order to make this comparison stable, we should find comparable siblings of those to by any stable internal complexity measure (maybe McCabe or Halstead is enough for now) and than compare the contributions (again by stable measures) over time

Best Frederik

Am 12.09.25 um 21:41 schrieb Enric Tobella Alomar:

Hi Frederik,

Thanks for laying out your thoughts so clearly.

I agree with the idea of experimenting before making real changes, but I think we need to be cautious with the assumption that moving from AGPL to LGPL automatically results in higher adoption and more contributions. We already have a couple of real-world “experiments” inside OCA itself:

- edi-framework
- queue_job

Both were licensed under LGPL rather than AGPL, and yet they did not attract significantly more contributors or maintainers compared to their AGPL counterparts. If anything, the contributor base has remained relatively small and fixed over the years. This suggests that the license alone is not the determining factor for contributions — other aspects like module complexity, required expertise, or the integrators’ business model also play a huge role.

So the licensing impact turns out to be limited (as our current examples suggest).

For me, it would be more relevant to study why communities like "Spanish Odoo Association" are able to attract so many supporters. They have a similar message, but they have a different strategy that allows to engage most of EE companies. Maybe these people are not making PRs, but at least they make a monetary effort that helps contributors.

Best regards,

El vie, 12 sept 2025 a las 21:22, Frederik Kramer (<notifications@odoo-community.org>) escribió:

Hi Raphael,

as always very detailed and very insightful thoughts. Honestly, i can't add much value here than just saying you are right with all you said in my opinion. The 20/80 relation for LGPL/AGPL sounds quite reasonable (even if Pareto edges almost always apply).

I'll take the fear / discomfort of Pedro (and Enric) very reasonable, so instead of doing to much to fast, I would suggest to start with a controlled experiment (that even Pedro and Enric would be willing to agree to). 

The experiment could look like as follows:

Take a small, but prominent baseline or infrastructure module that we know or assume many people use (even many in illegal ways (just like Tom pointed out) as of today) were a solid majority (of OCA members) and the whole responsible PSC believes would be better if it were licensed under LGPL (or at least has no objections). Lets the responsible owners induce the license change from AGPL to LGPL, advertise this change, make an effort to publish and post about the module, the change and its useful usage, encourage to actively contribute...

and than measure diversity, total amount, quality of contributions, speed of migrations etc. for that very module over a longer period of time (+/- 1 year) and compare it with the AGPL population of similarly reasonable baseline and infrastructure moduls licensed under AGPL. 

That way we can easily test hypothesis without taking much risk. If the most supported hypothesis (i.e. some few baseline / infra modules LGPL, 80% business logic modules AGPL -> induces more adoption / contribution) we should see first supporting data from that experiment.

Best Frederik

Am 12.09.25 um 17:23 schrieb Raphaël Valyi:

Hello, I think I need to share how I see the big picture.

But first, let me exemplify again with an Odoo market I know very well. You may know that OCA/l10-brazil is the most active OCA repo (14k commits, 4000+ PRs, 150k lines of code, 70 contributors). Not because Brazil is an ERP eldorado but exactly because it is often pointed as the hardest ERP market (you need 200+ tax fields on an invoice line, a company doesn't use all of them but certainly some 60-80, a diversity of 80% companies use may be 180 of them. Same thing e-invoicing has 800 fiscal fields and is over SOAP...).

Well there are now 50 official Odoo partners in Brazil, I'm pretty solid, the large majority is a scam of disposable noobs (half life = 1.5 year) who believed it was just about reselling Odoo EE. The vast majority just fail their projects like lemmings (they call us later) as soon as they venture outside of CRM or project management. As I follow the Github notifications I can guarantee you these 50 partners never contributed a SINGLE PR to the OCA. In fact it seems only people unable to do a line of code or use Google to get an overlook would partner. So the selection is pretty inverted (Dunning Krueger)... Instead, aside from Akretion, you now have Escodoo and Engenere who are serious people, CE only, and contributing many PRs to the OCA (outside of OCA/l10-brazil as well). But when I read this from Quartile https://www.quartile.co/en_US/blog/odoo-bits-pieces-1/essential-criteria-for-selecting-your-odoo-partner-as-an-end-user-company-120 let's say it matches my experience.

I also know the French market very well as I pretty much started it 15 years ago (remember openerp.com used the open source ERP whilepaper I wrote at Smile on their frontpage for some 3 years). And I can tell you the quality of the official partner network dropped a lot. 10 years ago they were a well intentioned elite (before Odoo turned it into a "market for lemon", and now, aside from a few exceptions they are mostly a scam, mediocre at best. Less than 5% of these French partners contribute PRs to the OCA on a yearly basis...

Overall, I feel Odoo is doing an unassumed transition from an innovative customizable ERP framework to a SaaS product. In fact they grew a bubble since the start. Since they had to rival with the $ 20 millions inflated Openbravo bubble, continuing with their 10x exaggerated SaaS business model from 2010 for Sofinnova (Fabien shared it with me, as the 2nd partner on the American continent, I helped convince Sofinnova, I protested to Fabien it was inflated but kept quiet as he suggested). Then came the "sorrySAP" crap in 2013, the invention of the millions of happy users worldwide, the $ 500 millions secondary market investments...

Odoo themselves raised little money (on the primary market), less than $ 15 millions I think. In fact, since the start that is the partner network who fueled the growth. Then Odoo "pivoted" and dumped the "stupid partners who believed in that free software concept", made all the impossible early cases possible, did a crazy R&D... Remember that the 1st TinyERP web client didn't come from TinyERP themselves but was a 3rd party contribution (by Axelor)...

In a way Odoo externalized the cost of the bubble to its partner network: "sign your project with the latest noob Gold partner who paid for its status and it will be like you will be doing your project with an Acsone or Akretion engineer with 10+ years of Odoo experience". Pretty much what they sold before Odoo EE was a product in itself...

It worked for a while. It grew in quantity while dropping in quality. This is exactly what is called a "market for lemon" with a quality converging to zero as it was shown by Nobel laureate George Akerlof

https://saylordotorg.github.io/text_introduction-to-economic-analysis/s19-01-market-for-lemons.html

At the same time Odoo has been improving its product a lot that is very True. Odoo is now quite well coded and is even pretty solid.

Finally, I think Odoo is in the middle of a transition: It is very likely Odoo Enterprise succeeds as a limited SaaS product for micro-companies (like Salesforce, Netsuite). Success will obviously depend on the country. And I think it's quite nice if they meet this success while funding the Odoo CE core under the LGPL license. Much like Basecamp or later Shopify funded the Rails framework.

What I find very "questionable" in fact is that they use the money from these partners they are fooling and their own customers to fund their transition toward an unassumed double agenda of a SaaS ERP for micro-companies. Indeed, Odoo will never assume the average quality of its partner network is crashing to zero.

But this is my vision: yes the partner network will stick to a very low quality for years and years (read again the implacable mechanics of the Market for Lemon) to come and an Odoo Enterprise code and license which is not designed for customizations or extensions but solely to protect the Odoo own IP.

And no, I don't see a bright future for this ecosystem of EE partners so that's why I suggest the OCA don't fool itself too much into trying to accommodate with the Odoo SA business roadmap.

And finally, while I said all this, I do share the concern that AGPL is a bit business unfriendly and I do agree we need some LGPL in the OCA to make it easy for companies using Odoo+OCA to protect some of their IP.

What is a good mileage? I don't know, maybe 20% LGPL and 80% AGPL would be nice.

@Tom:

About dual licensing again: it should be an opt-in option for the module authors but not forced otherwise you are simply hijacking the AGPL projection the modules authors might expect.

And also, the OCA will never be able to check if some business is using a valid LGPL exception module they purchased from the OCA. This simple fact would make it possible the AGPL would be violated massively meanwhile.

Thank you if you read it through ;-)

On Fri, Sep 12, 2025 at 1:57 PM Stéphane Bidoul <notifications@odoo-community.org> wrote:

Pedro,

Please don't assert things you can't possibly know about how other companies operate and why.

Best,

-Stéphane

On Fri, Sep 12, 2025 at 3:32 PM Pedro M. Baeza <notifications@odoo-community.org> wrote:

So you have provided the perfect example for confirming the hypothesis that going LGPL, the number of contributions will be reduced: how can it be that Tecnativa, having only 10 persons, contributes 4x more than companies like Camptocamp/Acsone, that have 40/50 persons?

They develop on top of enterprise modules, which they can't share, so they don't contribute to OCA.

They develop more private things, as they are allowed due to the license being LGPL, so they don't contribute back to OCA.

And again, remember the big vendor lock-in you are imposing on your customers installing enterprise modules, being the vendor Odoo S.A., not you. And even not advertising that to your customers (by ignorance or complacency). That's the big win of Odoo doing that the conversation doesn't turn around this.

Regards.

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

--

Raphaël Valyi

Founder and consultant

http://twitter.com/rvalyi

www.akretion.com

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

-- 
Dr.-Ing. Frederik Kramer
Geschäftsführer

initOS GmbH
Innungsstraße 7
21244 Buchholz i.d.N.

Tel:   +49 (0) 4181 13503 12
Fax:   +49 (0) 4181 13503 10
Mobil: +49 (0) 179 3901819

Email: frederik.kramer@initos.com
Internet: www.initos.com

Geschäftsführung:
Dr.-Ing. Frederik Kramer & Dipl.-Ing. (FH) Torsten Francke

Sitz der Gesellschaft: Buchholz i.d.N.
Amtsgericht Tostedt, HRB 205226
USt-IdNr.: DE815580155
Steuer-Nr: 15/200/53247

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

--

Enric Tobella Alomar

CEO & Founder

www.dixmit.com

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

-- 
Dr.-Ing. Frederik Kramer
Geschäftsführer

initOS GmbH
Innungsstraße 7
21244 Buchholz i.d.N.

Tel:   +49 (0) 4181 13503 12
Fax:   +49 (0) 4181 13503 10
Mobil: +49 (0) 179 3901819

Email: frederik.kramer@initos.com
Internet: www.initos.com

Geschäftsführung:
Dr.-Ing. Frederik Kramer & Dipl.-Ing. (FH) Torsten Francke

Sitz der Gesellschaft: Buchholz i.d.N.
Amtsgericht Tostedt, HRB 205226
USt-IdNr.: DE815580155
Steuer-Nr: 15/200/53247

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

--

Enric Tobella Alomar

CEO & Founder

www.dixmit.com

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe