TDF situation in relation to the German Foundations Authority, and why this led the BoD to make an extremely difficult decision on the MC election

italovignoli · October 3, 2024, 10:04am

This message is very long, because it attempts to explain in a simple way a very complex situation that goes back over the years to 2014, when the first tenders were launched, and has progressively worsened to its current severity.

In the past, someone has said that I am long-winded and verbose, which is only partially true as I am normally neither long-winded nor verbose except when I feel that the topic is complex and needs to be explained in order to be understood properly.

Often, and this is not a criticism but just an observation, people with a technical background tend to take a lot of information that is part of their professional heritage as taken for granted, whereas my experience – gained especially during the years I taught at the university – is the opposite, and suggests that topics should always be approached in all their facets, all the more so when they are very complex.

So, get ready for a detailed explanation of an extremely complex topic, perhaps one of the most complex I have ever dealt with in my professional life. And, if you see fit, thank my stubbornness, because even if I had settled for the brief summary made by lawyers, we would have continued to base all discussions more on hearsay than facts.

Preamble

Some facts:

In 2023, The Document Foundation was the subject of an Audit commissioned by the German Authorities that highlighted a number of problems that could result in anything from a “simple” fine to the loss of nonprofit status, with disastrous consequences for the LibreOffice project (imagine having to pay VAT on donations and taxes like a normal company, for multiple years, in addition to the image damage to open source software in general).
In this regard, Audit says: A withdrawal of the non-profit status could lead to high tax or other penalties. Based on the information available to us, it is not possible for us to determine with sufficient certainty whether the Foundation’s existing funds would be sufficient to make these potential payments.
As a result of the Audit, the German Authorities requested information from the three members of the BoD with a declared Conflict of Interest due to their affiliation with a company that, over the years, has had a commercial contractual relationship with The Document Foundation as a result of tenders for LibreOffice development or a free trademark license to sell the LibreOffice app on the Apple Store or Microsoft Store. One of these three members is Cor Nouws.
In this regard, Audit states: With regard to the contracts awarded to companies related to the Board of Directors, we were unable to obtain any arm’s length or third party comparisons. For this reason, we cannot conclusively assess whether these contracts would stand up to an arm’s length comparison.
With regard to the free provision of trademark licenses for app stores to companies related to the Board of Directors, we were not provided with any evidence that this service was appropriately compensated. Therefore, we cannot exclude a violation of the statutory use of the Foundation’s funds. In this context, we were not provided with sufficient evidence that there were no violations of the tax regulations for the recognition of non-profit status.
As a result of the Audit, the German Authorities have not yet made any decisions. This, of course, does not mean that the Audit has gone unnoticed, and that the risks associated with the Audit’s findings may not occur in all their severity.
In 2024, The Document Foundation is the subject of Audit. Obviously, it is the BoD’s task to demonstrate to the auditors that the results of the previous Audit have been given due consideration, and that the BoD has embarked on a path aimed at a progressive solution of the problems, once they have been analyzed and understood in their significant complexity. In this regard, after a series of discussions, including heated ones, the BoD unanimously reached agreement on a strategy that should lead to the resolution of the main problems by the end of the BoD’s current term.

Once you have defined the facts, before reading the rest of the document, I ask you the courtesy of clearing your mind of a number of potential preconceptions, of which I myself was a victim for a long time (until I decided to address the root of the problem by carefully rereading all the legal documents and discovering that a document crucial to understanding the issues had never been translated but only summarized by other lawyers, taking the content for granted, as if we all knew the rulings of the German Supreme Federal Court inside out):

Most of the problems highlighted are due to Conflict of Interest.
The Conflict of Interest is an inherently solvable problem, so it is not understandable why a “trivial” Conflict of Interest is valid and sufficient reason to invalidate the entire tendering process or the free grant of a trademark license, two decisions in relation to which at the time there was broad consensus among both BoD members and founders.

Also, it would be appropriate to avoid thinking that:

The companies in the ecosystem – and particularly allotropia and Collabora – have divergent interests from those of The Document Foundation. The reality is different, and there should be broad convergence on common interests, with the understanding that the companies have business interests that cannot be the same as a nonprofit foundation (only business interests).
The BoD has a specific interest in avoiding the election of a candidate in the MC because that way it can do whatever it wants. As if the bylaws are not of paramount importance, and they are bylaws that can only be changed with the approval of the majority of TDF Members. Perhaps it is time to go back to our basics.
AAA doesn’t want the tenders because it wants to have all the money in the bank, BBB wants the tenders to get rich, CCC is DDD’s friend and can’t stand EEE, FFF can’t understand and should listen to GGG, and so on. At this rate, we can use the whole alphabet for statements that serve no purpose, neglecting problem solving. We all, as individuals, make a percentage of our decisions wrong, and this is statistically demonstrable. Let’s start working together on problem solving.
Someone within the BoD has the majority of votes. Good ideas, and effective solutions have the majority of votes. There may be differences of view on some points, even substantial ones, but in the end what has to win is the project. If we keep thinking that we are against each other, for some reason, we will never get anywhere. Let us free our minds.

What is the real problem

Paragraph 8 of The Document Foundation’s Bylaws, “Tasks of the Board of Directors,” in Item (1) states among other things:

The Board of Directors is limited in its power of representation by the purpose of the Foundation.

I have a well-founded suspicion that no one has ever given much thought to this sentence. I certainly didn’t, and so – if I had ever been one of the BoD members who set up the tendering (procurement) process and granted the free trademark license to run the LibreOffice app in the Apple Store and the Microsoft Store – I would have unwittingly run into the “excess of power of representation,” and this – as a cascade – would have created the problems that we all face today. I emphasize the “all of us,” because if we fail to solve the problems, we risk finding ourselves with nothing left in our hands.

Thus, the real problem – the one that invalidates the awarding of tenders and related bills, and the free trademark license for stores – is the “excess of power of representation.”

A very complex problem, because it is not an absolute problem but a problem related to the object of The Document Foundation, which is made explicit in Paragraph 2 Item (1) of the Bylaws “Goals of the Foundation.”

The foundation aims to promote the following by supporting free software:

Public and professional education
Science and research, particularly in the field of computer science
Civic engagement for non-profit purposes

And of course to the status of a nonprofit foundation, which is made explicit in Paragraph 3 Item (1) of the Bylaws:

The foundation pursues exclusively non-profit goals as specified in the paragraph “Tax-deductible purposes” of the tax code.

At this point, the document in question comes into play, quoted and summarized (insufficiently in my opinion) by the opinion of a law firm specializing in foundations, which uses the May 27, 2021 Federal Supreme Court judgment to challenge the validity of the tenders and the free trademark license because the decisions by the BoD constituted an “excess of power of representation.”

A problem made even more serious by the presence of conflict of interest, but not specifically related to the latter. Unfortunately, the obsessive repetition of this term in email exchanges between members of the former BoD shifted everyone’s attention to this problem, to the point where it seemed to be the sole cause of all evil.

Admittedly, conflict of interest is much easier to understand than “excess of power of representation,” and this has contributed to this misunderstanding. I myself, even after reading all the documents once, kept wondering how it was possible that a “trivial” conflict of interest could be the cause of such significant problems, to the extent that it could jeopardize the nonprofit status of The Document Foundation.

Trying to simplify the concept, “excess of power of representation” occurs when the BoD of a nonprofit foundation makes a decision that is not or is not perceived to be equidistant from all stakeholders, and therefore is or can be interpreted as favorable to only one of the stakeholders.

The problem is solved by introducing the subtle concept of “arm’s length,” which obviously has nothing to do with arm’s length but refers to equidistance with all stakeholders. Thus, the BoD of a nonprofit foundation must be super partes, and in this, simply abstaining from a discussion or decision – as was, by custom, the case within TDF BoD – is not sufficient.

The Federal Supreme Court Judgment contains a couple of sentences that help to understand the specificity of the problem:

Paragraph 42: The distinction made by the Court of Appeal between a ‘narrow’ and a ‘broad’ purpose (which includes the criterion of charity) is not apparent. The appeal rightly points out that the limitation of the power of representation of the board to the ‘purpose of the foundation’ – which is sufficiently recognisable for legal transactions – is intended in particular to protect the foundation from entering into legal transactions that are detrimental to its status as a charitable foundation and the consequent risk of losing its recognition as a charitable foundation.

Point 43: It is true that in individual cases – as in the present case – it can be very difficult to assess whether an agreement fulfills the requirements for non-profit status under tax law. However, in view of the immense variety of possible constellations of legal transactions, it is generally not possible for charitable foundations to provide more precise rules. The non-profit rules are based on the fulfillment of certain tax-privileged purposes, which are defined in terms of content and not formally on the basis of certain legal transactions. Anyone wishing to enter into a contract with a foundation recognized as charitable must generally be aware that legal transactions detrimental to the charitable status will not be covered by the board’s power of representation. The trust placed in the foundation and the protection of legal transactions therefore do not justify the rejection of the validity of a statutory restriction on the power of representation of the foundation council, because the question of whether a contract is compatible with the status of a charitable organization may be difficult to answer in individual cases.

Thus, the “excess of power of representation” is a substantial problem, to the extent that it can jeopardize nonprofit foundation status, but it is most difficult to determine. According to the Federal Supreme Court’s Judgment, it is appropriate to take it almost for granted that any transaction that may be problematic for nonprofit foundation status – because it does not comply with the principle of equidistance to stakeholders – cannot fall under the BoD’s power of representation.

If we think about it, the history of the LibreOffice project – and its ancestor OpenOffice – makes it much more difficult to solve this problem, because over the years friendship, collaboration, and working relationships have been built that “bring closer” rather than keep the distance between stakeholders, and this in a way that is completely independent of the status of the individuals (volunteer, or employee, or collaborator of a company).

Personally, I believe that all the incidents in which the “excess of power of representation” occurred within The Document Foundation’s BoD occurred without the BoD members being aware of the problem, so much so that the practice of abstention had become customary, and up to a certain point – in 2020, when the BoD received its first legal opinion on the matter – it seemed the most obvious solution to the presence of Conflicts of Interest.

Unfortunately, and on this specific point the Federal Supreme Court’s Judgment leaves no doubt, abstention is not sufficient because it does not respect the principle of equidistance (arm’s length) among stakeholders, and on the contrary leaves room for the possibility of favorable treatment toward one (or more) of them. Thus, the history of the LibreOffice project, where two companies – allotropia and Collabora – participated in many software development-related tenders and got most of the contracts, makes the “excess of power of representation” occur in almost all cases.

Of course, the free trademark license only made this situation worse, because this decision also amounted to “excess of power of representation.” This, even in case the representatives of allotropia and Collabora had not been within the BoD, because the problem of “excess of power of representation” is independent of any kind of Conflict of Interest.

At this point, it is clear that the tender system adopted by The Document Foundation until 2020 is not sustainable as a procurement policy, as the “excess of power of representation” becomes an almost automatic consequence of any decision made on the basis of a common-sense system and not on a process structured in such a way as to preserve the principle of “arm’s length” based on legal criteria to ensure equidistance among stakeholders.

Carlo Piana is working on this “procurement policy.” Unfortunately, since the problem emerged – in relation to which there has been an excessive focus on conflict of interest, which in my view has skewed the perception and made conflict of interest the mother of all evils – perverse dynamics have been unleashed within TDF BoD (for which most members are responsible, each for his part) that have slowed the process.

At this time, it is completely useless to establish the weights of responsibility, because these do not change the problem, which is the blocking of the tendering process with an inevitable impact on the competitiveness of LibreOffice in the office suite landscape. And, of course, the crystallization of a situation of non-compliance with nonprofit foundation legislation, due to the recurrence of the “excess of power of representation,” which led to the conclusions contained within the audit.

Conclusions

To make it easier to understand the reasons that led the BoD to make an extremely difficult, and hopefully unique, decision, I will try to summarize the main points in sequence:

The Document Foundation’s 2022 budget was the subject of an audit, which found significant problems that could jeopardize the foundation’s nonprofit status and consequently its survival.
The authorities have not yet made any decisions with respect to the results of this audit, but they may do so at any time. In the meantime, it is a good idea for the BoD to take actions aimed at solving the problems highlighted by the audit, which the BoD is trying to do – with some effort – once the problems are understood and contextualized.
The Document Foundation’s 2023 budget is being audited in this precise time frame.
Cor Nouws was a member of the BoD in 2020-2022 and 2022-2024, which means that he is aware of the situation from the time of the first legal opinion (2020) forward. During the two terms, Cor Nouws declared his affiliation with Collabora, one of the two companies against which the BoD exercised the “exces of power of representation.”
The Membership Committee of The Document Foundation is the body responsible for coordinating the audit, deciding to whom to assign the task, and monitoring its performance. Because of its role, it may also be involved in the audit, at the discretion of the professionals assigned to the work.

Thus, the presence of Cor Nouws within the Membership Committee would have meant:

The simultaneous presence of the same person between the one who coordinates the audit and the one who is being audited, with all the possible consequences of the case under the law.
The repeated presence of the same person, who because of his previous affiliation is directly involved in the problems detected by the audit, and on whom the authorities have not yet commented, either in relation to the problems detected by the audit or in relation to the explanations provided by the person himself, within an official body of The Document Foundation. This, in contradiction to the need for change from the past highlighted by the audit.

Therefore, I described Cor Nouws’ candidature for the election of the Membership Committee as “irresponsible”, and confirmed my judgement by voting in favour of the motion of exclusion. I understand the extraordinary nature of this decision, but I continue to be astounded by the gross superficiality of the person who decided to submit his candidature in any case despite being aware of the situation.

timar · October 3, 2024, 10:43am

Whose idea was the first audit in 2023? Was it initiated by German authorities?

PaoloVecchi · October 3, 2024, 10:45am

Yes, it was initiated by the supervisory authority as TDF has reached a size that requires yearly audits.

thb · October 3, 2024, 11:05am

That sentence got added (the only change to our statutes, in 12 years), as a result of the board-initiated research & review of the situation. Which means, all decisions before are not necessarily constrained in the way you’ve outlined.

I’d therefore dispute your “no one has ever given much thought to this” statement rather strongly.

jfn · October 3, 2024, 11:29am

Hi,

For what it is worth, here’s a quick translation into French of Italo’s
message (translated by DeepL in private session, then very slightly
reviewed by myself). The translation is just that: a translation. Please
refer to the original Italo’s message for the more precise wording.

8< -------------------------------------------------------------------

Ce message est très long, car il tente d’expliquer de manière simple une situation très complexe qui remonte à 2014, lorsque les premiers appels d’offres ont été lancés, et qui s’est progressivement aggravée pour atteindre sa gravité actuelle.

Par le passé, quelqu’un a dit que j’étais prolixe et verbeux, ce qui n’est que partiellement vrai car je ne suis normalement ni prolixe ni verbeux, sauf lorsque j’estime que le sujet est complexe et doit être expliqué pour être bien compris.

Souvent, et ce n’est pas une critique mais une simple observation, les personnes ayant une formation technique ont tendance à considérer comme acquises beaucoup d’informations qui font partie de leur patrimoine professionnel, alors que mon expérience – acquise notamment pendant les années où j’ai enseigné à l’université – est inverse et suggère que les sujets doivent toujours être abordés sous toutes leurs facettes, a fortiori lorsqu’ils sont très complexes.

Préparez-vous donc à une explication détaillée d’un sujet extrêmement complexe, peut-être l’un des plus complexes que j’aie jamais traités dans ma vie professionnelle. Et, si vous le jugez bon, remerciez mon entêtement, car même si je m’étais contenté du résumé succinct fait par les juristes, nous aurions continué à baser toutes les discussions plus sur des ouï-dire que sur des faits.

Préambule

Quelques faits :

En 2023, The Document Foundation a fait l’objet d’un Audit demandé par les autorités allemandes qui a mis en évidence un certain nombre de problèmes pouvant aller d’une « simple » amende à la perte du statut d’association à but non lucratif, avec des conséquences désastreuses pour le projet LibreOffice (imaginez devoir payer la TVA sur les dons et les impôts comme une entreprise normale, pendant plusieurs années, en plus de l’atteinte à l’image du logiciel libre en général).
À cet égard, l’Audit déclare Le retrait du statut d’organisation à but non lucratif pourrait entraîner des pénalités fiscales élevées ou d’autres pénalités. Sur la base des informations dont nous disposons, il ne nous est pas possible de déterminer avec suffisamment de certitude si les fonds existants de la Fondation seraient suffisants pour effectuer ces paiements potentiels.
À la suite de l’audit, les autorités allemandes ont demandé des informations aux trois membres du conseil d’administration ayant déclaré un conflit d’intérêts en raison de leur affiliation à une société qui, au fil des ans, a eu une relation commerciale contractuelle avec The Document Foundation à la suite d’appels d’offres pour le développement de LibreOffice, ou d’une licence de marque gratuite pour vendre l’application LibreOffice sur l’Apple Store ou le Microsoft Store. Cor Nouws est l’un de ces trois membres.
À cet égard, l’Audit déclare : En ce qui concerne les contrats attribués à des sociétés liées au Conseil d’administration, nous n’avons pas pu obtenir de comparaisons avec des tiers ou avec des parties indépendantes. Pour cette raison, nous ne pouvons pas évaluer de manière concluante si ces contrats résisteraient à une comparaison de pleine concurrence.
En ce qui concerne la fourniture gratuite de licences de marques pour les magasins d’applications à des sociétés liées au conseil d’administration, nous n’avons reçu aucune preuve que ce service était rémunéré de manière appropriée. Par conséquent, nous ne pouvons pas exclure une violation de l’utilisation statutaire des fonds de la Fondation. Dans ce contexte, nous n’avons pas reçu de preuves suffisantes qu’il n’y a pas eu de violations des règles fiscales pour la reconnaissance du statut d’organisme à but non lucratif.
À la suite de l’audit, les autorités allemandes n’ont pas encore pris de décision. Cela ne signifie évidemment pas que l’audit est passé inaperçu et que les risques associés aux conclusions de l’audit pourraient ne pas se concrétiser dans toute leur gravité.
En 2024, The Document Foundation fait l’objet d’un audit. Il est évident qu’il incombe au CA de démontrer aux auditeurs que les résultats de l’audit précédent ont été dûment pris en compte et que le CA s’est engagé sur la voie d’une résolution progressive des problèmes, une fois que ceux-ci ont été analysés et compris dans leur grande complexité. À cet égard, après une série de discussions, y compris des discussions animées, le conseil d’administration est parvenu à un accord unanime sur une stratégie qui devrait conduire à la résolution des principaux problèmes d’ici la fin du mandat actuel du conseil d’administration.

Une fois les faits définis, avant de lire la suite du document, je vous demande la courtoisie de vous débarrasser d’un certain nombre d’idées préconçues potentielles, dont j’ai moi-même été victime pendant longtemps (jusqu’à ce que je décide de m’attaquer à la racine du problème en relisant attentivement tous les documents juridiques et en découvrant qu’un document crucial pour la compréhension des questions n’avait jamais été traduit mais seulement résumé par d’autres avocats, en prenant le contenu pour acquis, comme si nous connaissions tous les arrêts de la Cour suprême fédérale d’Allemagne à la perfection) :

La plupart des problèmes mis en évidence sont dus au conflit d’intérêts.
Le conflit d’intérêts est un problème intrinsèquement soluble, et il n’est donc pas compréhensible qu’un conflit d’intérêts « insignifiant » soit une raison valable et suffisante pour invalider l’ensemble de la procédure d’appel d’offres ou l’octroi gratuit d’une licence de marque, deux décisions qui, à l’époque, faisaient l’objet d’un large consensus parmi les membres du conseil d’administration et les fondateurs.

Il convient également d’éviter de penser cela :

Les entreprises de l’écosystème – et notamment allotropia et Collabora – ont des intérêts divergents de ceux de The Document Foundation. La réalité est différente, et il devrait y avoir une large convergence sur les intérêts communs, tout en comprenant que les entreprises ont des intérêts commerciaux qui ne peuvent pas être les mêmes que ceux d’une fondation à but non lucratif (uniquement des intérêts commerciaux).
Le CA a un intérêt spécifique à éviter l’élection d’un candidat au sein du CM, car il peut ainsi faire ce qu’il veut. Comme si les statuts n’étaient pas d’une importance capitale, et ce sont des statuts qui ne peuvent être modifiés qu’avec l’approbation de la majorité des membres du TDF. Il est peut-être temps de revenir à nos fondamentaux.
AAA ne veut pas des appels d’offres parce qu’il veut avoir tout l’argent à la banque, BBB veut les appels d’offres pour s’enrichir, CCC est l’ami de DDD et ne supporte pas EEE, FFF ne peut pas comprendre et devrait écouter GGG, et ainsi de suite. À ce rythme, nous pouvons utiliser tout l’alphabet pour des déclarations qui ne servent à rien, en négligeant la résolution des problèmes. En tant qu’individus, nous prenons tous un pourcentage de décisions erronées, ce qui est statistiquement démontrable. Commençons à travailler ensemble à la résolution des problèmes.
Au sein du conseil d’administration, quelqu’un dispose de la majorité des voix. Les bonnes idées et les solutions efficaces ont la majorité des votes. Il peut y avoir des divergences de vues sur certains points, même importants, mais en fin de compte, c’est le projet qui doit l’emporter. Si nous continuons à penser que nous sommes opposés les uns aux autres, pour une raison ou pour une autre, nous n’arriverons jamais à rien. Libérons nos esprits.

Quel est le vrai problème ?

Le paragraphe 8 des statuts de The Document Foundation, intitulé « Tâches du Conseil d’administration », au point (1), stipule notamment ce qui suit :

Le Conseil d’administration est limité dans son pouvoir de représentation par l’objet de la Fondation.

Je soupçonne à juste titre que personne n’a jamais réfléchi à cette phrase. Je ne l’ai certainement pas fait, et donc – si j’avais été l’un des membres du CA qui ont mis en place le processus d’appel d’offres (approvisionnement) et accordé la licence de marque gratuite pour faire fonctionner l’application LibreOffice dans l’Apple Store et le Microsoft Store – j’aurais involontairement rencontré « l’excès de pouvoir de représentation », et cela – en cascade – aurait créé les problèmes auxquels nous sommes tous confrontés aujourd’hui. J’insiste sur le « nous tous », car si nous ne parvenons pas à résoudre les problèmes, nous risquons de nous retrouver sans rien dans les mains.

Ainsi, le vrai problème – celui qui invalide l’attribution des appels d’offres et les projets de loi y afférents, ainsi que la licence de marque gratuite pour les magasins – est l’« excès de pouvoir de représentation ».

C’est un problème très complexe, car il ne s’agit pas d’un problème absolu, mais d’un problème lié à l’objet de The Document Fondation, qui est explicité au paragraphe 2, point (1) des statuts « Buts de la Fondation ».

La fondation a pour but de promouvoir les éléments suivants en soutenant les logiciels libres :

L’éducation publique et professionnelle
La science et la recherche, en particulier dans le domaine de l’informatique
L’engagement civique à des fins non lucratives

Et bien sûr, le statut de fondation à but non lucratif, qui est explicité au paragraphe 3, point (1) des statuts :

La fondation poursuit exclusivement des buts non lucratifs tels que spécifiés dans le paragraphe « Buts déductibles » du code des impôts.

C’est alors qu’intervient le document en question, cité et résumé (insuffisamment à mon avis) par l’avis d’un cabinet d’avocats spécialisé dans les fondations, qui s’appuie sur l’arrêt du Tribunal fédéral du 27 mai 2021 pour contester la validité des appels d’offres et de la licence de marque gratuite parce que les décisions du CA constituaient un « excès de pouvoir de représentation ».

Un problème rendu encore plus grave par la présence d’un conflit d’intérêts, mais qui n’est pas spécifiquement lié à ce dernier. Malheureusement, la répétition obsessionnelle de ce terme dans les échanges de courriels entre les membres de l’ancien CA a focalisé l’attention de tous sur ce problème, au point qu’il est apparu comme l’unique cause de tous les maux.

Certes, le conflit d’intérêt est beaucoup plus facile à comprendre que « l’excès de pouvoir de représentation », et cela a contribué à ce malentendu. Moi-même, même après avoir lu une fois tous les documents, je me suis demandé comment il était possible qu’un conflit d’intérêts « insignifiant » puisse être à l’origine de problèmes aussi importants, au point de mettre en péril le statut d’organisation à but non lucratif de The Document Foundation.

Pour simplifier le concept, on parle d’« excès de pouvoir de représentation » lorsque le conseil d’administration d’une fondation à but non lucratif prend une décision qui n’est pas ou pas perçue comme étant équidistante de toutes les parties prenantes et qui, par conséquent, est ou peut être interprétée comme favorable à l’une d’entre elles seulement.

Le problème est résolu en introduisant le concept subtil d’« arm’s length », qui n’a évidemment rien à voir avec la longueur d’un bras mais se réfère à l’équidistance avec toutes les parties prenantes. Ainsi, le conseil d’administration d’une fondation à but non lucratif doit être super partes, et en cela, le simple fait de s’abstenir d’une discussion ou d’une décision – comme c’était, par coutume, le cas au sein du conseil d’administration de TDF – n’est pas suffisant.

L’arrêt du Tribunal fédéral contient quelques phrases qui permettent de comprendre la spécificité du problème :

Paragraphe 42 : La distinction faite par la Cour d’appel entre un but « étroit » et un but « large » (qui inclut le critère de la charité) n’est pas évidente. Le pourvoi souligne à juste titre que la limitation du pouvoir de représentation du conseil d’administration au « but de la fondation » – qui est suffisamment reconnaissable pour les actes juridiques – vise en particulier à protéger la fondation contre la conclusion d’actes juridiques préjudiciables à son statut de fondation d’utilité publique et le risque qui en découle de perdre sa reconnaissance en tant que fondation d’utilité publique.

Point 43 : Il est vrai que dans des cas individuels – comme dans le cas présent – il peut être très difficile d’évaluer si un accord remplit les conditions pour obtenir le statut d’organisme à but non lucratif en vertu du droit fiscal. Toutefois, compte tenu de l’immense variété des constellations possibles d’opérations juridiques, il n’est généralement pas possible pour les fondations d’utilité publique de fournir des règles plus précises. Les règles de non-lucrativité reposent sur la réalisation de certains objectifs fiscalement privilégiés, qui sont définis en termes de contenu et non pas formellement sur la base de certains actes juridiques. Toute personne souhaitant conclure un contrat avec une fondation reconnue d’utilité publique doit généralement être consciente du fait que les actes juridiques préjudiciables au statut d’utilité publique ne seront pas couverts par le pouvoir de représentation du conseil d’administration. La confiance placée dans la fondation et la protection des transactions juridiques ne justifient donc pas le rejet de la validité d’une restriction statutaire du pouvoir de représentation du conseil de fondation, car la question de savoir si un contrat est compatible avec le statut d’une organisation caritative peut être difficile à trancher dans des cas individuels.

Ainsi, l’« excès de pouvoir de représentation » est un problème important, dans la mesure où il peut mettre en péril le statut de fondation à but non lucratif, mais il est très difficile à déterminer. Selon l’arrêt du Tribunal fédéral, il convient de considérer comme presque acquis que toute opération susceptible de poser problème pour le statut de fondation à but non lucratif – parce qu’elle ne respecte pas le principe de l’équidistance des parties prenantes – ne peut relever du pouvoir de représentation du CA.

À bien y réfléchir, l’histoire du projet LibreOffice – et de son ancêtre OpenOffice – rend ce problème beaucoup plus difficile à résoudre, car au fil des années se sont construites des relations d’amitié, de collaboration et de travail qui « rapprochent » plutôt que de maintenir la distance entre les parties prenantes, et ce de manière totalement indépendante du statut des individus (bénévole, ou salarié, ou collaborateur d’une entreprise).

Personnellement, je pense que tous les incidents dans lesquels « l’excès de pouvoir de représentation » s’est manifesté au sein du CA de The Document Foundation se sont produits sans que les membres du CA soient conscients du problème, à tel point que la pratique de l’abstention était devenue habituelle, et jusqu’à un certain point – en 2020, lorsque le CA a reçu son premier avis juridique sur la question – elle semblait être la solution la plus évidente à la présence de Conflits d’Intérêts.

Malheureusement, et sur ce point précis l’arrêt du Tribunal fédéral ne laisse aucun doute, l’abstention n’est pas suffisante car elle ne respecte pas le principe d’équidistance (arm’s length) entre les parties prenantes, et laisse au contraire la place à la possibilité d’un traitement favorable à l’une (ou plusieurs) d’entre elles. Ainsi, l’histoire du projet LibreOffice, où deux sociétés – allotropia et Collabora – ont participé à de nombreux appels d’offres liés au développement de logiciels et ont obtenu la plupart des contrats, fait que l’« excès de pouvoir de représentation » se produit dans presque tous les cas.

Bien entendu, la licence de marque libre n’a fait qu’aggraver la situation, car cette décision constitue également un « excès de pouvoir de représentation ». Et ce, même si les représentants d’allotropia et de Collabora n’avaient pas fait partie du conseil d’administration, car le problème de « l’excès de pouvoir de représentation » est indépendant de toute forme de conflit d’intérêts.

À ce stade, il est clair que le système d’appel d’offres adopté par The Document Foundation jusqu’en 2020 n’est pas viable en tant que politique d’approvisionnement, car l’« excès de pouvoir de représentation » devient une conséquence presque automatique de toute décision prise sur la base d’un système de bon sens et non d’un processus structuré de manière à préserver le principe de l’« arm’s length » basé sur des critères légaux pour assurer l’équidistance entre les parties prenantes.

Carlo Piana travaille sur cette « politique d’achat ». Malheureusement, depuis que le problème est apparu – à propos duquel on s’est trop focalisé sur les conflits d’intérêts, ce qui, à mon avis, a faussé la perception et fait du conflit d’intérêts la mère de tous les maux – des dynamiques perverses se sont déclenchées au sein du CA de TDF (dont la plupart des membres sont responsables, chacun pour sa part), qui ont ralenti le processus.

À l’heure actuelle, il est parfaitement inutile d’établir les poids de responsabilité, car ceux-ci ne changent rien au problème, qui est le blocage de la procédure d’appel d’offres avec un impact inévitable sur la compétitivité de LibreOffice dans le paysage des suites bureautiques. Et, bien sûr, la cristallisation d’une situation de non-conformité à la législation sur les fondations à but non lucratif, en raison de la récurrence de « l’excès de pouvoir de représentation », qui a conduit aux conclusions contenues dans l’audit.

Conclusions

Pour faciliter la compréhension des raisons qui ont conduit le CA à prendre une décision extrêmement difficile et, je l’espère, unique, je vais essayer de résumer les principaux points dans l’ordre :

Le budget 2022 de The Document Foundation a fait l’objet d’un audit qui a révélé des problèmes importants susceptibles de mettre en péril le statut d’organisation à but non lucratif de la fondation et, par conséquent, sa survie.
Les autorités n’ont pas encore pris de décision par rapport aux résultats de cet audit, mais elles peuvent le faire à tout moment. Dans l’intervalle, il est souhaitable que le CA prenne des mesures visant à résoudre les problèmes mis en évidence par l’audit, ce que le CA s’efforce de faire – avec un certain effort – une fois que les problèmes sont compris et contextualisés.
Le budget 2023 de The Document Foundation fait l’objet d’un audit à ce moment précis.
Cor Nouws était membre du CA en 2020-2022 et 2022-2024, ce qui signifie qu’il est au courant de la situation depuis le premier avis juridique (2020). Au cours de ces deux mandats, Cor Nouws a déclaré son affiliation à Collabora, l’une des deux sociétés contre lesquelles le CA a exercé l’« excès de pouvoir de représentation ».
Le Comité des membres de The Document Foundation est l’organe chargé de coordonner l’audit, de décider à qui confier la tâche et de contrôler son exécution. En raison de son rôle, il peut également être impliqué dans l’audit, à la discrétion des professionnels chargés de ce travail.

Ainsi, la présence de Cor Nouws au sein du comité des membres aurait signifié :

La présence simultanée de la même personne entre celui qui coordonne l’audit et celui qui est audité, avec toutes les conséquences possibles de l’affaire au regard de la loi.
La présence répétée de la même personne, qui, en raison de son affiliation antérieure, est directement impliquée dans les problèmes détectés par l’audit, et sur laquelle les autorités n’ont pas encore fait de commentaires, que ce soit par rapport aux problèmes détectés par l’audit ou par rapport aux explications fournies par la personne elle-même, au sein d’un organe officiel de The Document Foundation. Cela va à l’encontre de la nécessité d’un changement par rapport au passé, soulignée par l’audit.

Par conséquent, j’ai qualifié la candidature de Cor Nouws à l’élection du Comité des membres d’« irresponsable » et j’ai confirmé mon jugement en votant en faveur de la motion d’exclusion. Je comprends la nature extraordinaire de cette décision, mais je reste stupéfait par la superficialité flagrante de la personne qui a décidé de présenter sa candidature en tout état de cause, bien qu’elle ait été informée de la situation.

-------------------------------------------------------------------- >8

Best,

italovignoli · October 3, 2024, 11:59am

I have just provided my opinion, based on the information I have. I am happy if someone has though about the issue, and added that missing sentence to the statues. On the other hand, I continue to think that the issues related to the “power of representation” are not known by the majority of TDF Trustees, and have been overlooked at least by several BoD Members. This is not an accusation, but a simple observation. The issue is so complex and peculiar that I think it is absolutely normal to ignore it, unless you are deeply involved in the governance of a non-profit foundation and care deeply about all the legal implications.

ohallot · October 3, 2024, 12:12pm

I have translated to Brazilian Portuguese and published in

italovignoli · October 3, 2024, 12:29pm

I would like to add some background to my post:

I have drafted the text on my own, without sharing with other BoD members (apart from informing them that I was writing it), based on research and analysis that I have carried out on my own because I did not feel comfortable with the general perception about the issues raised by the Audit.
The post represents my own opinions, which may or may not be shared in part or in their entirety by other members. So, I am not representing the whole BoD, and - most important - I have not written the post based on a common strategy of the BoD members who have expressed the same vote.
I have decided to write the post because I was feeling uncomfortable about not providing a full explanation of my vote, as I completely understand that there may be completely different - and perfectly legitimate - opinions about my vote.
I have a background in humanities, which means that when I face a challenge - and, believe me, the current situation at TDF is a major challenge - I start from studying it in detail, before developing my own ideas about it. Of course, reading several times dozens of legal documents (and translating other legal documents which were still in German) in order to have the full picture in front of me, took quite a substantial chunk of my available volunteer time. I would have loved to provide the information at an earlier stage, but it was not possible. From now on, though, I will be more responsive.
If someone reads my post as a personal accusation, he is wrong. Of course, there is a specific personal accusation to Cor Nouws, which is rather clear, and has been explained in details. Cor and I probably see our roles inside the BoD and/or the MC in a completely different way, but having different opinions is rather normal when you are part of a complex organization.
I would like to see the factions disappear. We have internal problems to solve, and external problems to address (if we do not want LibreOffice to gradually fall into oblivion). The solutions in some cases are complex, but they are within our grasp, and only if we can work together on the basis of common goals (forgetting personal goals for a moment) we can achieve them.

Is all from me for today. Now back to day-to-day work.

PaoloVecchi · October 3, 2024, 2:52pm

Thanks Italo for the excellent summary.

As you might be aware I’ve been working on this specific topic in June 2022 together with our legal counsel, Caolan and Emiliano, during the short period of time when the legal subcommittee has been allowed to work on various legal matters that you mentioned, which then eventually resulted in the decision to amend TDF’s statutes.

The analysis I made of the judgement at the time, and reviewed again following your message, is for the best part aligned with the summary of your conclusions with a slight difference in interpretation for which we could ask advice so that we can work with the same understanding of the matter.

It is true that most of the focus has been put on conflict of interests and violation of the arm’s length principle as for several years these were the indicators that something wasn’t quite right based on commonly applied rules and regulations. We discussed this at length within the board and in this forum about the behaviour that should be shown by directors with personal interests that can lead to conflict of interests. As some still have different ideas on how to handle their personal interests and avoid them to become actual conflict of interests we should work on clarifying that once and for all.

In my reading the judgement goes beyond the issue of conflict of interests and the violation of the arm’s length principle as it’s clear that these issues are indicators that these legal transactions, and the behaviours leading to approve these transactions, might be contrary to the purpose of the foundation and/or detrimental to the charitable status with the result that the power of representation has been exceeded.

As from the judgement: “Anyone wishing to enter into a contract with a foundation recognised as charitable must generally be aware that legal transactions detrimental to the charitable status will not be covered by the board’s power of representation.”

With a very short and non exhaustive explanation we could say that the issues we are facing are in good part due to decision taken in conflict of interests and due to violations of the arm’s length principle which could lead to potential loss of charitable status so those decision and the resulting actions exceed the power of representation of the board.

There are surely other situations where the power of representation might be exceeded so directors will need to pay extra care during the whole decision process and in case of doubts consult our legal counsel.

So, IMHO, the judgement does not represent something new but is a good clarification of what happens in situation when certain common principles are not respected.

hmatsuura · October 3, 2024, 3:25pm

Muito obrogado @italovignoli pela explicação e como membro, apoio a decisão tomada pela diretoria, com base nas robustas justificativas estando, muito bem explicado e fundamentada.

EyalRozenberg · October 3, 2024, 4:08pm

I want to thank @italovignoli for emphasizing some of the legal principles which (formally) govern the foundation, and the BoD’s conduct and the application of its powers. I also think this is a good framing of the problematic aspect of the TDF’s relations to ecosystem companies. However, the description of this situation as a “problem” is only part of the truth. Because much, or perhaps even most, of the LibreOffice development work is done by ecosystem companies and their employees; and a large fraction of the trustees are employees or affiliates of these companies. And with the foundation being democratic, then, naturally, some candidates will either be from these companies or closely associated with them. If this is a “problem”, than the solution must be the tearing away of body of people from the TDF; not a huge body, but a significant and active one. If we were just designing the TDF today from scratch, and if we had enough active trustees, I’m guessing I would support stronger statuory guarantees against our symbiosis with ecosystem companies. But right now we can’t define the flesh of our flesh as a “problem”, and assume that nautrally, they should not be members of the BoD or MC because there’s a change this might be inconvenient or create problems for us.

Something like your post could be a start of a process of discussion of whether we want to have a symbiotic TDF as our basic form, and accommodate legal restrictions as best we can (e.g. special independent and separate TDF body for tendering); or whether, on the contrary, we want to switch the “arms-length” variant of a foundation, in which case we encase the interaction with ecosystem company employees and affiliates away from decision-making.

But that process of strategic deliberation is not one which the BoD can hold itself, private, consider held and completed, then piggyback its conclusion - which is not shared by the trustees - onto a retroactive disqualification of an MC candidate, after not having brought any of this up before and during the elections.

Comments:

Cor would naturally have been excluded from the auditing process. If you don’t have faith that this would be the case, then you (= the BoD) should have at least brought this matter up after his candidacy had been presented, vis-a-vis all candidates, and in the town halls.

It’s not a repeated presence, it’s a different presence, in a different body, and a non-excecutive one. Also, you are asking us to assume implicitly that the previous board, and Cor particularly, has engaged in gross misconduct which merits a ban on him serving in official capacities in the TDF. While many may share that belief, the TDF has never established this to have been the case, and you can’t just do so suddenly and with zero due process. Like you said, “the authorities have not yet commented” on Cor’s conduct as a BoD member. Alternatively, you may be arguing that it should be an impersonal moratorium on Collabora/Allotropia affiliates; but then, what about @lpranam and @bvarga91 ?

It’s more than just that. A BoD member can abstain from voting and still be extremely influential on the vote: By bringing up the issue and promoting a discussion and the taking of the decision; by their influence on other board members; and by earlier influence (even if legitimate in itself) on the composition of the BoD.

Questions:

What about internal actions within the foundation, rather than transactions with an outside party? Can those be null/void, or nullable/voidable, if determined to be detrimental for the nonprofit status?

Is this audit taking place now? Will it take place after 2024 is concluded?

Was this strategy published and I missed it? Was the process and discussion for reaching this agreement published?

Importance in principle is one thing, and being binding in practice is another thing. I don’t believe the BoD acts as if bound by the bylaws in practice; especially because our mechanisms for enforcing the bylaws on the BoD are barely existing to being with, and atrophied. And also with the MC being the subject of the current row.

PaoloVecchi · October 4, 2024, 12:34pm

That is incorrect. “The board” did not initiate any research and review of the situation, the legal subcommittee that proposed the vote did the research and review. Then from the board only the then chairperson engaged in questioning the matter and only in relation to the effect it would have on, at the time, current and future contracts with suppliers.

The judgement has been released in 2021 but it covers events that started in 2011.The decisions found to be not in line with the purpose of the foundation and/or detrimental to the charitable status, as mentioned by Italo and reported in legal statements and the audit, also in my opinion did exceed the power of representation. To be noted that legal statements and the audit already found some of these decisions to be invalid and the excess of power of representation was additional to them.

Italo is mostly right. As written above there has been a brief discussion, between the then chairperson, Emiliano and myself, about the urgency of the vote we sent out and how that sentence would affect current and future contracts. Apart from that back and forth no other elements of the judgement have been taken in consideration by the majority of the board including when it has been explicitly communicated that some decisions being taken, after we implemented that sentence, were very likely exceeding the power of representation of the board. As the majority of the previous board did have issues in respecting the arm’s length principle and avoiding influencing discussion in which they were in conflict it was clear that most did not give much thought to this matter.

thb · October 4, 2024, 2:42pm

Hi Paolo,

thx for the additional details, and indeed as you said, a not insubstantial part of the previous board did spend quite some time on the matter.

One remark:

Indeed the board did, but it was initiated even earlier (I’m talking about the entire review of the situation, around what, how & where TDF is structuring its activities - of which the change to the statutes was a partial result). To set the record straight: it was @uwealtmann and me, later joined by @lothar.becker , who were first hunting for experts, then starting the conversation with the Winheller law firm.

As I said, I’m disputing this topic was hardly ever considered on the boards I’ve served.

mmeeks · October 4, 2024, 3:38pm

Hi Italo,

I am not a lawyer, and I am confident that TDF has its own counsel to protect and structure its activities, so I will not wade into the detail here on my own. I’m grateful that you pointed out that there should be broad convergence on common interests between TDF and the ecosystem.

The write-up also seems to me, while interesting, to be not that relevant to whether the statutory election process has been followed fairly in Cor’s case.

For a more complete and helpful perspective on tendering I posted this some time ago and re-link it - which puts some numbers around this:

It is also worth bearing in mind that with the average full-time staff salary, as of some years back when I was on the board, many individual staff have been paid more in total than any software development supplier.

Beyond that; I again make again the offer I’ve made before - if competitive tendering on the open market (something that is not generally required anyhow for TDF) is in some way not a good enough way to get a fair market price (how?) - I would be well up for getting a 3rd party audit of our pricing, time logging & margin (if any) there. That has been an open offer met with silence for a long time AFAIR.

Ultimately there is always the suspicion that the lack of focus and urgency to tackle these issues, is part of a political agenda to disenfranchise corporate trustees (and now even ex. corporate trustees which is new).

I understand that for your part you appear to want to fix things rapidly, which is great! once again - I’d recommend talking - I want the dream that was LibreOffice/TDF to rock and am open to helping constructively.

Thanks !

Michael.

EyalRozenberg · October 5, 2024, 9:23am

I must stress that, while your post explains your motivation and perceptions, it does the opposite of explain or legitimize your vote, i.e. you focused on the TDF’s overall course and on the conduct of past BoDs and Cor as a director - and did not establish: 1. The basis for a claim of illegitimacy / disqualification of Cor’s candidacy (as opposed to whether it is desirable or not). 2. What gives the BoD the power to disqualify a candidate retroactively.

italovignoli · October 5, 2024, 10:47am

I respect your opinion, but the reality is that I provided plenty of reasons for rejecting Cor’s election. Of course, you may disagree with my opinions, but they are explained in details in my post.

italovignoli · October 5, 2024, 10:54am

I have never written that ecosystem companies cannot participate in tenders. On the contrary, having a procurement policy written by a lawyer - a sound procurement policy can never developed following on a discussion amongst TDF members, as the result would probably be the same as the former procurement policy, which was developed following that kind of discussion - will allow all stakeholder to bid, including ecosystem companies. Of course, solving the issues based on the arm’s length concept will not be easy, but is definitely possible, and this is what Carlo Piana is doing.

EyalRozenberg · October 5, 2024, 2:43pm

You’ve provided reasons why you don’t want him / think it’s a bad idea for him to be in the MC - but not the two things you (and the BoD) need to provide, which I listed: A basis for the claim of Cor’s illegitimacy as a candidate, and the source of BoD power to retroactively disqualify him. On the contrary, the reasons you provided seem to be rooted in your interpretation of the foundation’s best interests (which I may not disagree with); yet, someone’s candidacy being a bad idea or detrimental to the foundation in certain ways - even if that is conceded as the objective truth - does not make him disqualified, nor grant the BoD the power to disqualify them. Your decision still appears as being “ultra-vires”, even had it been taken before the elections, and after them - it gains another aspect of illegitimacy which is your failure to disqualify earlier.

Even someone completely agreeing with all your reasons must still demand that you recognize you did not have grounds for disqualification and could not disqualify retroactively; that you acknowledge Cor’s election as an MC member; and finally, if you believe you have both grounds for his removal and the power to remove him (e.g. as per §8.4) - that you make that claim explicitly. I doubt that you can establish such a claim, and thus will all simply have to “live with” Cor as an MC member for the next term (assuming our statutes don’t change).

thb · October 5, 2024, 3:41pm

Hi Italo,

again thx for the very extensive, independent & (likely) time-consuming summary of the issue!

Beyond my earlier (rather tangential) remark, I’ve encountered two more questions, that I don’t see answered (neither in your initial positing - and I’ve read it several times in full! - nor in any of the follow-ups):

That is indeed true, and in my view it is a feature rather than a bug (for any OpenSource community, not just for LibreOffice). But as you rightly state, this community-of-peers setup has saddled us with some challenges, on how to structure our governance.

My first question now is: how are the past issues with the arm’s-length-principle, while ecosystem companies had representatives on the board, any different from the current situation? Right now, there are four board members with significant portions of either their personal, or company income, originating in TDF. If I understand your rationale for disqualifying Cor in the MC election correctly (violation of the arm’s-length-principle) - would not the board need to act similarly also for this case? Certainly there’s collaboration, and a working relationship, perhaps even friendship, amongst directors?

My second question relates to this paragraph:

Are you sure, that merely “demonstrating […] due consideration, and […] embarking on a path aimed at […] solutions” will be enough for the current year? My reading of the audit result was, that plausible explanations are required, that past contracts and licenses were not in violation of relevant tax regulations?

In case the board would want to take that route, I’d of course repeat my earlier offer (similar to @mmeeks ’ one), that a suitable 3rd party would get access to allotropia books and timelogs, to ascertain that indeed the tenders we’ve worked on were not overpriced.

Best, Thorsten

cornouws · October 5, 2024, 7:51pm

Hi Italo,

Referring to what I wrote before:

… let me do that here.

Thanks again for your massive work, Italo, in sharing this information that, to my knowledge is mostly correct!
Then: alas I don’t know/understand why items that were full on the boards agenda at the end of last year are still there, but trust people’s best effort of course.

Additional I notice more useful information to understand the situation (in a broader context) is missing. Apparently pieces that you don’t know or have missed it in the massive amount of communication (and no one made you aware). But since you, we ,are looking for understanding, let me add these. (Briefly now; in much more detail later.)

The audit having taken place, and the remarks in the report, are correct. I think that is is important to realize that legal report, that is at the base of the audit, could have been significantly better. It has been pointed at that the author of the legal statement could have had better information about some of the important processes in TDF. Some fundamental questions made to the statement, related to topics that ended in the audit, could have been looked at prior to the audit. From the audit itself, it seems more information could have been (pro-actively) shared with the auditor. Internal, I would say it is useful to make the discussion around place/role of (members that are working for) ecosystem companies not needlessly complicated, by avoiding to pick only part of the legal advise. As already mentioned in other comments: offerings to resolve at least one of the issues, at that time did not lead to action. For these reasons it looks to me that a ‘less troublesome picture’ of ecosystem companies and a much better audit for TDF could have been possible.
I can understand trustees will be interested to learn more of this; hence more later.

So short: I do agree that taking care for CoI’s and arm-length principle are important (and not new) but need to add that these are really not the (only) problems.

Cheers,
Cor