[Archive Papers Privacy Free Software as a Community of Mutual Aide]
Increasingly, the rights and capabilities to own technological artefacts, where they exist at all, are re-served to corporations and not to citizens. There are historical, economic, metaphysical, ideologicaland cognitive reasons for this situation, in addition to purely technological factors, which we will traceby following the fate of various concrete examples, analysed into five categories of ownable elements.These categories are those of ownable function, ownable expression, ownable data, ownable installa-tions and ownable economies. In this paper, we attempt to align these goals of ownability into a research and activism program by describing a set of revolutionary goals in each category, and tracing ways wecould reach them
There are a lot of the ideas and Ethics proposed here that line up with The Complete Computing Environment, and the Arcology goals and desires.
The authors present a "digital serfdom", where various important classes of artefacts, increasingly essential for everyday life, including participating in political and economic life, cannot be effectively owned by ordinary individuals.
Consider iliana's ISP Wobscale Technologies, an effort that costs her thousands of dollars and is a considerable part of her hobby budget. The running cost of having IP addresses on the public internet, along with the benefits those can be used for, are a more-than-part-time hobby for an individual. I believe that there is room for collective action here, of course, and collective action seems to be the path forward in any way. This is the conclusion that the authors reach as well.
In this paper, we use “ownership” to refer to a set of related use patterns where owners can pursue alternatives, repurpose and adapt their property, or maintain their property as the surrounding world changes. We will survey the kinds of things that we might want to own and how we could set abouttransforming our environment so that they could be owned.
Ownership of Function
In talking about "ownership of function", the authors wonder why the color pickers between application frameworks must be reimplemented and cannot be shared. A simple example, but one that is true of my life, still1, as a user of both GTK and Qt free desktop applications. And it's those sorts of incompatibility and forced context-switching barriers that drive my adoption of Evil Mode and many of the thoughts expressed in the Principles of Complete Computing and in the Goals of the Arcology.
In a broader example, a contributor wants to extend iNaturalist to support interesting statistical analyses and curate more types of information, or simply to provide a list of species which have been confirmed in a free-form map area. For this to be implemented this contributor, a third party uninvolved in the development of iNaturalist, has to work out a way to share functions between overlapping communities, and to unify "multiple, local, partial perspectives." This is an issue of ownership where federation and decentralization can provide aide, allowing for organic growth of protocols and featuresets. But for these monolithic deployments, even if the model data is available, there is a community which is now split between platforms, unless one completely consumes the other.
Ownership of Expression
in expressing "ownable expression", there is the idea of being able to "buy into" and "buy out of" the expressions of others, and overlay your own expressions on top, sharing the new system as freely as its inputs:
We haveformulated this requirement as theOpen Authorial Principle: “Any expression by one author can haveits effect replaced by anadditionalexpression by a further author” (Basman et al., 2018) When we saythat this “should be possible”, we imply that it should be possible in practice — that is, that the authors’expressions can be successful at a cost which is affordable. This is a statement as much about economicsas it is about engineering. These costs can be measured in numerous terms, for example, investment in aparticular skill-set, direct monetary terms, effort in building a relationship with a particular community,or cognitive effort.
Importantly, the user owns the cost of this:
We note here that while some heroic efforts might succeed in transplanting this harmless-looking colourwidget, the resulting assembly would be little more than a curiosity — the hapless user would effectivelyhave taken on the responsibility for maintaining their own personally customised version of the entire application.
I've expressed frustration in the past that software engineering is far too young of a profession for how highly it holds itself. The authors reference work from 1988 decrying our "pre-industrial environment", and little feels like it's changed from this as a developer, if not a user, we're simply blind to the issue and costs
The fact that the user has to be troubled by these irrelevant considerations of platform betray that we arein a pre-industrial environment plagued by irregularities like screws that are not of a size to match screwthreads, trains that cannot run on each other’s tracks, or electricity that may not travel between differentwire networks (David and Bunn, 1988).
These issues exist across architectural and technical bounds, programming languages constrain us to the points where they can be integrated, the authors point to the idea of an "integration domain" which programming languages can fit in to and integrate together freely. This feels a lot like WASM, web assembly:
The ultimate implication of these issues is that programming languages as they are currently constitutedneed to be abolished, through a process of being dissolved into an “integration domain”, a constructdescribed by Kell (2009). An integration domain is a space in which one aligns multiple artefacts toco-operate, and which, importantly, is the only space that may contain assumptions about the particulars of several artefacts. This implies that artefacts to be integrated expose structured representations ofthemselves (Clark and Basman, 2017).
Anyone inducted into the Eleusinian mysteries of code takeson board as implicit knowledge that they will have to acquire a broad spectrum of “jungle skills” whichhelp them navigate these kinds of informational swamps — not least, an in-depth study of the structuringconventions used by the particular technology in question, but more broadly applicable techniques suchas deliberately banging on things to see what breaks, studying the structure of long stack traces, orsearching across the codebase for the occurrence of tell-tale strings
I have noticed a lack of information process more generally, but navigating unfamiliar codebases is an acute example. people unable to effectively use google search, is another expample, but this inability to solve for yourself was a constant issue i noticed within the engineering culture at Uber. Indeed, as an effective learner, and with years of cultural knowlege, i was seen as a sage, nearly all-knowing in certain domains and through certain lenses.
Ownership of Data
I've began pondering on this in data portability is not enough and it's been one of the anchorpoints of Moving Towards Real Privacy, but natural monopolies and centralised services are simply incapable of providing meaningful data ownership, especially when the data is about a citizen rather than having been created by this citizen. They may not even be aware of its existence, and unable to be readily informed of it (Informed Consent is a canard).
The edge that this paper adds is that it's cost on many axes that prevent it, some of them are issues that can be mounted by some users, but only on plateaus of privilege2.
In building systems which allow for self-ownership on top iNaturalist as the returning, or any other open source centralized community, there are options with unfavorable economics.
We can expose a power-user-like view of the database and say "go to town." This is a model we used at Uber in the earlier days when we focused on empowering our "on-the-ground" teams to make business decisions of their own, and the main tool that they had access to was a simple web-based SQL executor. Regional and City Managers, Operations etc, could run analytics queries to fulfill marketing campaigns, run targeted campaigns and provide data to regulators that allowed the business to run, and in the process created an unaccountable system of surveillance that took years to be locked down to the point where it wasn't a vector for spying and stalking, and even longer for the central offices to provide tools which replaced those thousand-line adhoc queries.
This provides a flexible, but ultimately unkind interface, and doesn't necessarily solve for real-time collaboration which is increasingly seen as a required feature by consumers and power-users alike. CRDTs and OTs come close but the UX is still bad, there's still a lot of thought that has to go in to and come out of the usage of these datatypes and others like Purely Functional Data Structures.
but at the end of the day, these integrations rely on us being able to own our data, right? if we have all these malleable tools and even good UX for managing, but nothing that can fit in to them what is the point? Open data, and processes for managing and versioning data sets for reproducability is a wide concern within the scientific community. Efforts towards this are Open Knowledge Labs with their Frictionless Data initiative, and the Wikimedia foundation.
The authors of this paper keep returning towards the idea of "lenses" as a computing construct, composable, layerable systems which add, remove, and modify data and its representation. By building systems this way, they can be applied to remote data sources to bring them in to expected form, a form portable. This is very close to the ideas in some of the Hypergoal interpretations Arcology: a computer which takes streams of remote data, and can store, present, and process that information in ways which are valuable to me.
The way data flows through systems, both open and closed, is opaque still. This is something that I learned at Uber again, where most of my experience with data systems resides: finding a table in Hive with user data in it, knowing nothing about where the data came from and who it's gone to, or who even owns it. Mystery data, years old, unquestionably valuable and valueless.
It would not be possible for a user to trace the lineage of aparticular piece of data to discover how and when it had been introduced, or what previous versions it appeared in, without a costly manual procedure. The process of data version management needs to beas ownable as anything else — with the possibility to “take away” entire historical lineages of data with their sequencing and provenance relationships intact.
Tagged data can in-advertantly be solved with technologies like Datashards, I believe, and many folks believe certain classes of Blockchain implementations are perfectly suited for this task, with the expected UX guarantees and costs thrown out the window. I think it's important to find a solution which is not reliant on computationally expensive solutions to an philosophers' problem. This is a problem which industry faces3 and I believe they will solve data-tagging and lineage tracking in the long-term, eventually gifting The Commons through some open source implementation.
This is about the complexity of running an ISP as I mentioned at the beginning, or any sort of highly reliable infrastructure. For software, the authors contend that this is closer to solved than any of the other ownership issues what with kubernetes and openstack and configuration automation pushing the cost of deployment towards 0. YunoHost is trying to do this for individuals and small organizations and I think this is a very promising direction, but "turmoil" in the ecosystem is still a big issue and laymen usability is still far off.
This is a highly promising development, although the ecosystem is still in turmoil and is very far from the state where an ordinary citizen (or expert) could turn up and at the push of a button set up a fully configured incarnation of their own version of a complex platform such as iNaturalist.
In evaluating solutions towards the current incarnation of the Arcology and this wiki system, I researched Knowledge Management tools that are, basically Small Technology. I found that NoteSelf syncs with CouchDB, and there were a number of other TiddlyWiki solutions like one which used. A lot of the ideas in the Small Technology Foundation overlap, and indeed these factions would be happy allies. authors call out Hoodie as another promising primitive.
Now we get to the utopian thinking… without much more thought than i have put in to, the authors put forth cooperative organizing as a drive to a lot of the political and social ethos barricading the ownership changes outliend. as a reaction to Uber and the rest of the gig-economy shysters and Surveillance Capitalists. woops! I've often joked, with more than a little truth to it, that Uber Turned Me in to an Anti-Capitalist and I guess I'm not the only one.
The Platform Cooperativism Consortium's Platform Cooperative Development Kit is good prior art and something to keep in mind when considering a A Job Worth Keeping.
they've laid out a blueprint for radical progress, a blueprint I've myself intuited over the course of my growth as a technologist. it's a radical future, communities of mutual-aide forming around loose standards and tools, blurry at the edges. an emacs sub-community for everyone, if you will. free software's bizarre bazaar, but with equitable structure to the core.4
we're talking revolution now, baby. rise up rise up.
but i don't disagree with them.
The references in this would be a gold mine of expansion in my Privacy and ethical reasoning.
Firefox can use KDE file dialogs↩︎
I pulled this phrasing out my ass, but maybe it's useful.↩︎
Tech is Not Ready for New User Rights↩︎
The Cathedral and the Bizarre↩︎