Attending the Decentralized Web Summit at the Internet Archive wasn’t a bad way to spend my first week living in San Francisco. Big thanks to Peter Van Garderen and Courtney Mumma for encouraging me to go.
It was an exciting, challenging and inspiring few days. There are some excellent reports on the overall programme out there; including Brewster Kahle’s, Mouse Reeve’s or Maira Sutton’s. There’s also a slew of media reporting from Wired, NYtimes, Fortune, Boing Boing and more. So rather than reviewing the event as a whole I’d like to try to highlight some of the technologies, projects and ideas that struck me as important for recordkeepers to know about. I’ll also (tentatively) make some suggestions about how our professional practices might fit into this emerging world.
Archiving the Web
A decentralized Web is inherently archive-friendly. It’s open in the sense that its protocols are available to all, and it’s highly redundant, in that it replicates content across the globe. Projects like the Inter Planetary File System (IPFS) have produced an alpha which embodies these principles. IPFS is a peer-to-peer hypermedia protocol that aims to supplement, or possibly even replace, the Hypertext Transfer Protocol. It means websites and web apps with no central origin server so they will persist regardless of single site failures or retirements; historic versioning (like git); and resilient networks for mirroring of data.
Compared with Web archiving tools like the Wayback Machine or PANDORA, which crawl the Web, take snapshots and then store the content with a single organisation, an IPFS- based Web would have qualities of permanence, comprehensiveness and robustness against censorship or attack. All this is interesting to consider in terms of appraisal when it is done with a view to making retention / deletion decisions; this self archiving web removes the notion of retaining some content longer than other content, or selecting content for long term retention. Everything is automatically part of the archive, from creation. The question of removal or redaction in response to breaches of personal privacy is also interesting to consider in this context. The creators of IPFS and other working in this space are looking at mechanisms in the publishing process to allow for these contingencies.
Data set provenance and governance
DAT is another really exciting project, which focuses on publishing and management of datasets using a decentralized model. This is sorely needed, as so many of the current systems for sharing of datasets currently fall down on ensuring persistence, keeping authoritative versions and recording provenance. DAT embraces the same core principles as IPFS, in that it is open and decentralized and ‘historical’, meaning that data is versioned with cryptographically secure hashes. It is then possible to computationally prove the data is the version it purports to be and to audit the history of the data.
Decentralized Autonomous Organizations: Next gen recordkeeping
The best example of what decentralized autonomous organizations could mean for our understandings of recordkeeping is perhaps the Ethereum Democracy DAO. In this explanation we see how it is possible, for a given community, to embed its requirements into code to ‘do business’ (voting, executing commands, paying people, turning IoT devices on and off) with full transparency, complete accountability and complete immunity from any human interference. While the network lives, the contracts will execute exactly the code they were created to execute, without any exception, forever. The code keeps the records of the execution of all this work, and it cannot be lost or deleted. The recordkeeping is the business – and as we recordkeepers know, this is optimal for authenticity.
DAOs can be implemented for any group to do almost anything – from governments to activist groups. A DAO could be implemented by an archive to work with record creators to ensure the capture and management of records, whatever their format. Here we would see interfaces with other tools for keeping information, perhaps something like IPFS. Peter Van Garderen wrote about this kind of idea in his piece ‘Decentralized Autonomous Collections’ earlier this year .
Any DAO will, however, require decision making about what objectives, values and principles its creators want their code to embody, through to decisions about the data that is kept recording each transaction. To explore some interesting thinking on high level functional requirements for these types of organisations, I recommend Christopher Allen’s Revised “Ostrom’s Design Principles for Collective Governance of the Commons”.
In addition to such governance questions, anyone establishing a DAO will also need to understand the nature of its (to quote Chris Hurley) ‘documents, deeds and doers’, their authority and relationships, over time. These are some of the areas in which I think recordkeepers’ skills come into play:
- identifying requirements and risks associated with making a record – in defining smart contracts (whether for a democracy DAO or otherwise), decisions to make a transaction (and therefore a record) should be made with consideration of the business, legal and wider societal context, the requirements and the risks that apply. The analysis techniques of appraisal (the more holistic use of the term as used elsewhere on this blog and as explained in ISO 15489:2016) apply here.
- identifying and tracking change in agents, functions, types of records/transactions, and mandates – in the DAO, like any recordkeeping instance, is made up of all of these elements – but they are not static. The people who are members of the group change over time; the business done by the group changes, as do the rules for that business. Keeping records of these that permit ‘roll back’ to the people, rules and business types that applied at a given point in time, including their relationships with each other, is core recordkeeping work. This needs translation into this new environment. Here, perhaps, is an opportunity for a new recordkeeping application to support DAOs?
I think it is exciting that with the emergence of these technologies we have the opportunity to focus our efforts as recordkeepers in a more concentrated way – the need to expend effort and resources worrying about immutability, permanence and availability being substantially reduced. There are no doubt many more potential opportunities for us recordkeepers and archivists to contribute to the growth of the decentralized web and blockchain enabled business other than the couple I have touched on here. This post won’t be the last attempt I make to explore these.
You can watch videos of keynotes and other stuff from the summit via the Internet Archive.
 Peter explains in the introduction to his piece: ‘A Decentralized Autonomous Collection is a set of digital information objects stored for ongoing re-use with the means and incentives for independent parties to participate in the contribution, presentation, and curation of the information objects outside the control of an exclusive custodian.’