Let's develop peasant computing, let's breed “kittens”
Framasoft + AMIPO
In recent years, we have witnessed the widespread concentration of internet practices among a very limited number of online service providers, represented by what is now known as GAFAM (Google Apple Facebook Amazon Microsoft). This centralisation, which is totally contradictory to the origins of the Internet, which was conceived as decentralised and distributed 1, led Tim Berners Lee, the creator of the web, to formulate proposals for the future 2. So, why should we be concerned about this? Simply because data, and particularly our personal data, are the economic fuel of these major actors, and such an accumulation of information about us gives them immense power, turning us into “products” thanks to their “free” services. The questions this raises are many and complex: generalised surveillance, artificial intelligences fed by “big data”, the end of anonymity and private life, brakes on freedom of expression and access to information, censorship or loss of data following the closure of a service… Fortunately, a band of irascible Gauls, meeting around the Association Framasoft 3 is bravely trying to “de-googlize” the Internet 4 and extend this initiative so that we can “retake control” 5.
These services track us everywhere, while claiming to give us a better “user experience”. But our behaviour is under constant surveillance. This information can be used to display targeted adverts, but the revelations of the Snowden case have also shown that Internet giants have been forced to communicate this data (sometimes extremely private: emails from Gmail, photos shared on Facebook, Skype conversations, smartphone locations, etc.) to the authorities. Under the pretence of fighting terrorism, states are able to gather much more intelligence than “Big Brother” could ever have dreamed of.
Our data is an extension of ourselves. It tells third-parties where we are, who we are with, our political and sexual orientations, sites we have visited, our favourite recipes, our favourite topics of interest, and so on.
While a single data point is not always sensitive, the loss of large amounts of aggregated data can be dangerous (for example if you browse topics about cancer before subscribing to a life insurance).
Your private life is an essential part of your individuality, and in a world where everything has been digitized (e-books, TV, phones, music, social networks, etc.), it would only take a malicious hacker with access to your smartphone a few minutes to cause you serious harm (taking control of your identity on Facebook, consulting your professional or medical information, making purchases without your authorisation, etc.).
The major actors of the Internet have become real giants: Google owns YouTube and Waze, Facebook has acquired WhatsApp and Instagram, Microsoft distributes Skype, etc.
This concentration of actors creates multiple issues: what if Facebook were suddenly shut down? And how could we browse the Web if Google went down? We rely more and more on services provided by a small group of suppliers. For example, Apple (iPhone), Google (Android) and Microsoft (Windows Phone) dominate almost the entire mobile OS industry.
Furthermore, the size of these actors impedes innovation: it’s hard to launch a startup that can match up to Apple or Google (the first and second worldwide market capitalisations, respectively).
Finally, The lack of diversity of the giants means they can track many people who are unaware that there may be alternatives, and it can influence the kind of data you receive (a Google search will produce different results for the term “nuclear power” depending on whether Google considers you to be an environmentalist or pro-nuclear power).
Web services used on your computer, smartphone, tablets (and other devices) are usually hosted on the “cloud”: servers spread across the planet, that host not only your data (emails, pictures, files, etc.), but also the application code.
For your data, this raises the issue of sustainability (what would become of your files if Dropbox were to close tomorrow?) and of your ability to switch easily between services (how would you recover your data from Facebook or Picasa and import it, with all the adjoining comments, into another service?).
For applications, this means that you are completely at the mercy of your service provider when it comes to proliferation of advertisements, changes to the user interface, etc., and that you have hardly any control over the way an application works. It is a “black box” that can exhibit malicious behaviour (sending spam SMS without your knowledge, executing malicious code, and so on).
In short, these companies trap us in gilded cages: gilded yes, but cages nonetheless!
“De-googlize” the Internet
Framasoft, through the “de-googlize” the Internet initiative, wishes to counter these threats to our digital lives by offering free, ethical, decentralised, and solidarity-based services. They are making a list of the most-used proprietary software, linking them to the corresponding free software they offer for those same services 6.
In 2017, around forty online services were offered free to internauts with a view to meeting a variety of needs: cloud-type personal file storage, calendars, contacts, collaborative document editing, video conferencing, cartography, mind mapping, meetings and surveys, distribution lists, social networks, online books, search engines, educational games, project management… the list is long, but “the way is free”.
The story of the Internet itself is one of free software, and this goes for standards as well as protocols. Its potential and popularity are a cause for envy, and large companies would like nothing better than to control it by imposing closed-source, locked-down, and non-interoperable systems.
For the Internet to stay true to its founding principles, those which have led to its success, users must be able to choose free software. That is to say, software whose source code remains open and accessible and is covered by a free software license.
Framasoft is committed to only using software with “free” source code, promoting an internet that allows exchange and independence.
Framasoft undertakes not to exploit its users’ data, and to promote a fair and open Web.
Through the services we deploy, we promote an economic model based on sharing costs and resources, and providing widespread access.
This model also has an educational aspect because we believe that by documenting ways to setup services, many users will in turn be able to share these resources.
We think that, by not infantilizing users and by sharing responsibility for the use of services, it will be possible to regulate abuse.
Framasoft is thus committed to promoting respect and autonomy for its users (as long as this is reciprocated).
To ensure equality for all, whether citizens or businesses, not only is it essential to avoid monopolies, but large organizations must be prevented from grabbing personal or public data.
Using tutorials to explain how to increase the use of free solutions that will allow a fairer Internet, we help to distribute codes and diversify usage.
Framasoft is thus committed to facilitating self-hosting and interoperability, so that its users don’t get “locked in”.
The K.I.T.T.E.N.S. project9
In the light of the success of their “de-googlize” the Internet campaign, Framasoft has seen a vertiginous increase in the number of users of their online services, with a corresponding increase in the work needed to maintain and guarantee those services, without ceasing to propose new ones. The Association, lead by five permanent members, relies largely on donations and does not want to grow beyond “human” size.
To extend the dynamic and encourage the decentralisation of services, Framasoft therefore launched the KITTENS initiative, with the aim of bringing together different structures and initiatives hosting services, data and content in their own way, but respecting a common manifesto and charter 10. Both documents are collectively written and modified by the members of the collective, to take into account the evolution of the different structures and the technical, social and legal context of data hosting in France.
The KITTENS collective employs a model of governance directly inspired by Free Software. Decisions concerning the evolution of the collective and the charter are taken in a collegiate fashion. Like source code, the collective model can be duplicated and modified by whoever wants to adapt it, for example, to specific regional contexts.
Each member is invited to participate in collective decision making by consensus, as far as is possible. In the case of conflicts of opinion, decisions are made by a simple majority vote.
The domain chatons.org is maintained and hosted by Framasoft (as long as that is possible and until the collective decides something else). It is made up of a website, with a list of members, and also a distribution list that enables members to communicate and exchange. They are invited to collaborate in the creation of public content for the site, to inform the public of information relating to KITTENS and its free hosting.
KITTENS has no administrative statutes as it principally consists of a public list which determines the members and a set of documentation to facilitate the exchange of knowledge, capitalisation on good practice, and information sharing.
Any organisation that respects the principles of the current manifesto and the KITTENS charter can propose itself as a member. In order to become a member, the collective must receive contact information for the organisation and at least one member of the organisation must be subscribed to the discussion list. Apart from the discussion and possibly some advice, there may also be a vote by simple majority to decide on the the acceptance of a new organisation into the collective.
One or various members can reserve the right to request the expulsion of another member, so long as the following conditions are met:
- the proposal must be supported by convincing arguments that are shared with all the members;
- it must be accepted by a collective vote, with or without counter arguments.
In awareness of the fact that it is not possible to guarantee respect for all the points in the KITTENS charter without threatening the confidentiality of the personal data held in the member's systems, peer control will de facto be imperfect. The collective therefore relies on trust and on the benevolence of the agreements reached between members.
KITTENS should therefore meet among themselves, respecting each others’ points of view, and find good practices and rules for inclusion, for questioning or expelling a member, prioritising respect for fundamental freedoms and the private lives of the users of collective services.
Each member, referred to below as a "KITTEN", commits herself to obey the charter that defines the following principles:
Transparency, non-discrimination and protection of personal data.
Honesty is the watchword of these commitments, which seek to establish the reliability of the proposed services and build user trust. The General Use Conditions should be perfectly clear, accessible and non-contradictory with the KITTENS charter.
The host should implement a transparent user account management policy, without discrimination, whether access is free or paid for. It must respect the jurisdiction of the country in question.
The host commits to allow all users to recover their personal data, encrypted or not, except in the case of particular services based on the the momentary transfer of encrypted personal data.
Openness, economy, protection
The services proposed should meet some technical requirements. Notably, servers should be based on free software solutions. Such software will make the reproducibility of the service possible, without generating additional developments in the structure of the server, or as a contribution to free software.
The use of open formats is obligatory, at least for all data sent to users. This is a clear policy in favour of interoperability. When the use of open formats is impossible, the data should be under a free license and be available for as many operating systems as possible. The sources should be made accessible.
Members of KITTENS commit to respecting the free licenses for the software that they use (which includes mentioning those licenses and referencing the sources).
In ethical terms, sponsoring is acceptable, as is patronage, donations, or having an economic model that consists of demanding payment for some functions or even for the entire service. The economic model of each KITTEN should be clearly expressed on a dedicated page that is easy for users to consult and understand. Evidently, the economic aspects of the activity of any KITTEN should strictly conform to the legislation of the country in question.
On the other hand, no advertising coming from advertising agencies will be accepted. Likewise, there should be no exploitation of personal data, there will be no monitoring of user activity except for legal and statistical ends, user addresses may not be used for anything other than administrative or technical ends. Statistical tools should be free and meet the Collective's conditions.
Solidarity and dissemination
KITTENS members owe each other assistance and mutual aid, through a discussion list or any other means, including periodic meetings. That is how KITTENS members can progress their services. One of the most effective means for maintaining systematic mutual aid is by contributing to the development of free software tools.
Nevertheless, members should not keep to themselves as that would only satisfy a limited number of people and would create discrimination in access to services. On the contrary, all efforts at communication with the public are encouraged as a way of disseminating free hosting solutions and creating links around the principles defended by the collective. The means should be shared, through trainings, public information sessions, stalls during demonstrations, speeches at conferences, the publication of leaflets, etc.
KITTENS services cannot be hosted by an actor who, by reputation, does not promote the neutrality of the Net. Data packages should be transmitted by KITTENS services without discrimination, which means the content, source or destination should not be examined. No communications protocol can be given priority in the way information is distributed. And no data can have its content arbitrarily altered.
The neutrality of KITTENS is therefore a political neutrality, as the convictions of members will be neither examined nor sanctioned, so long as they keep within the framework of current applicable laws.
AMIPO, an experience of a KITTENS construction in Orleans
AMPIRO, (“L’Association de Maintien de l’Informatique Paysanne Orléanaises,” the Association for the Maintenance of Computing for Farmers in Orleans) is a part of the French “AMAP” (Association of Organic Agriculturists who provide their fruit and vegetables directly to subscribers), with the idea of finding good bytes in a local association, rather than on the American “supermarket”. A “farmer” is defined as anyone who works for self-sufficiency and contributes to developing the environment and the countryside
Signing up to a local KITTENS initiative starts with bringing together a group of people motivated by the idea of practical reflections about the why of the how. From the first prototypes of services installed on recycled servers behind a decidedly asymmetrical internet uplink (with more download than sending capacity via ADSL 11), they aim to pass to the “production” phase on servers hosted with ethical providers in synchrony with our base values (Such as ARN 12 in Alsace or Tetaneutral 13 in Toulouse).
That requires setting up a legal entity (in this case, a collegiate association 14, without a president or head office); opening a bank account; organising a launch event to raise funds; creating content to raise awareness and setting up workshops for popular education; agreeing about the technologies used and the services proposed; deciding whether the welcome page of the website should be in http or https; creating the necessary communication and documentation tools; organising the installation and administration of the servers; making sure we are in accordance with the charter, particularly putting in place the necessary safeguards; proposing times and channels for communications in order to respond to user's questions…
The main aim of AMPIRO is to offer a personal cloud service (based on NextCloud) that allows inventories of files, contacts and calendars, for free, with the possibility of having more space by joining the association. The enthusiasm of the collective does not stop there, there may be a VPN (Virtual Private Network 15), or work on end-to-end encryption, so that we cannot read the data that are trusted to us, in order to be able to propose services to local associations or to accompany cooperative projects in the IT usage.
With our calloused hands, as crude farmers in computing, we wish to plant as many seeds as possible in the heads of our fellows so that little pixelated kittens can run about on free and tree-lined roads.
1. In the first instance, in order to convince CERN that a system of global hypertext was interesting enough for the research centre, this document foreshadows the World Wide Web as we know it today: https://www.w3.org/History/1989/proposal-msw.html ↩
7. Benjamin Bayart: Internet libre, ou Minitel 2.0? https://www.youtube.com/watch?v=AoRGoQ76PK8 ↩
8. Translator's note: Minitel was a centralised pre-Internet videotext terminal and service in France: https://en.wikipedia.org/wiki/Minitel ↩
9. Translator's note: The French initiative is called “C.H.A.T.O.N.S.” (https://chatons.org/). Chatons means “kittens” in French, and stands for “Collectif des Hébergeurs Alternatifs, Transparents, Ouverts, Neutres et Solidaires” (the Collective of Alternative, Transparent, Neutral and Solidarity-based Hosters). This is translated into English as “K.I.T.T.E.N.S.” (Keen Internet Talented Teams Engaged in Network Service). For more information see: https://framagit.org/framasoft/CHATONS/blob/master/docs/Charter-en.md ↩
12. http://arn-fai.net/ ↩