A what wiki?

From federated to distributed collective maybe, but what does this point to in terms of supporting human collective collaboration?

I didn't write yesterday, because what I envisioned in my last entry, two days ago, came about much more quickly than I had anticipated. I had a good conversation with John in the morning, and he pointed me again to Ward's session where he discusses ideas and developments around the Fedwiki. So I went along, expecting to sit quietly in the background, but far from it, I was warmly and delightfully welcomed into the conversation.

Having heard the conversation, and followed up watching some of the videos on Marc Pierson's vimeo channel (he was also there), I can share a few things that weren't obvious to me at first glance about the Fedwiki. Marc has also very generously offered to give me a guided introduction to Fedwiki, and I will know more after that. But I need also to mention my inspiring conversation with Revathi this morning, and more with Ria and Helen at home, without which I don't think I would have been prompted to put these words on the screen.

So here are my first impressions, though these may not reflect the reality of Fedwiki. The thing that sticks in my mind most strongly is that Wikipedia has become part of the established order of things. It now has traditions, and is in a very tangible sense traditional. It remains a commons, in the sense that it is governed by (some of) its users, though there are a large majority of users – just readers – who play no part in the editing or governance. And perhaps it is common for traditional commons to feel like institutions, with clear roles, governance practices, norms and conventions.

Perhaps it was this aspect that Ward was reacting against, in deciding that a new approach to wikis was needed. I connect this (once again) to Robert Kegan's work. While Wikipedia can be seen as having settled into Kegan's 3rd order of consiousness, the world of tradition, the Fedwiki pushes through to Kegan's 4th order, where people claim their own self-authorship. The idea is that every Fedwiki user has their own pages, which are editable only by them, but the system seamlessly skips across different authors' pages. This is enlivened by the fact that any user can copy (called ‘forking’, a term from collaborative software writing) a page, and make their own additions or alterations, under the same page title. There can therefore be any number of Fedwiki pages with the same title, being edited by different people. And note also that, just as anyone can make a copy of and edit your page, so you can copy back their changes to your page, if you want to.

If you then get a rich array of pages on one and the same topic, how to choose? There is an algorithm, whose details I know nothing of as yet, which selects one that is likely to satisfy you. We could imagine several inputs to that algorithm. Some might be based on the page ‘graph’ – that is, the details of which pages link to, or are based on, which others. Ones that have been more often borrowed from might be seen to be more valuable. Other inputs may be more personal. I could imagine being directed to a page that is more closely linked to my pages, for instance. I can also imagine other inputs, which are not yet feasible, based on personal ratings of pages. “People like you preferred this page” I can imagine some software calculating, or “people who liked the same pages as you preferred this page”. I'm deliberately evoking the collaborative filtering techniques that are so familiar to users of Amazon shopping sites.

There are some other very neat features already built into Fedwiki. (I say ‘already’, because the software is continuing to develop.) One feature is to help you see the differences between two pages. Another that I believe is there, though I haven't actually seen it yet, is that the system helps you trace the original author of any passage. I'm looking forward to learning more.

There are many things I am still dubious or puzzled about (looking forward to clarity), but the main positive that I can see remains that Fedwiki moves on from Kegan's 3rd to his 4th order of consciousness, in terms of the supporting web technology. It seems like the epitome of public self-authorship. By that very fact, though, it is not yet well adapted to collectivity. Collectivity can be 3rd order, of course. As noted above, Wikipedia covers that ground, but in a way that does little for co-creation, generativity or emergence. How might we envisage new technology offering those qualities that Wikipedia doesn't support? That's what I really want to know, to meet my vision of the potential of a knowledge commons.

Even with established wiki technology, I can still imagine generativity behind the scenes. If a small group of editors know and trust each other, they can pool their knowledge and abilities, and harvest their collective intelligence, bringing to life more than any of them could alone. Of course, that would be very hard to scale up to the size of Wikipedia. But it could happen, indeed I have suggested that we could try it out, on the P2P Foundation wiki. Individual author/editors would be members of small teams, and pages would be curated by teams, not by individuals. However, there is nothing in established wiki technology that contributes to good collaboration of that kind.

So here is what really excites me about all this. We could start with Fedwiki as it is, and we could, collectively (maybe needing trial and error) work out what is needed, progressively, to support collectivity, in the sense of collective intelligence, yes, but also collective motivation, engagement, sense of ownership, sense of community, and in some senses of the word, identity.

This process would be dealing with the technical features, but it could work as a development in parallel with human consciousness, moving collectively beyond Kegan's 4th order. I mean, in doing the technical development, looking at new features and what underlies them, we might just be able to glimpse what might be going on when people make that inspired jump from self-authorship into the fluid world of self-transcendence.

I'm now back with the what? again. Let this be an invitation to any reader who gets a sense of the potential of what I'm trying to write about, to be in touch.

Topics: Knowledge commons; Wiki software

If you have any remarks on any of my posts, please send me e-mail, saying what you want me to do with your remarks. Are they private to you and me, or would you be happy to quote you (I will always attribute your words unless you ask me not to), and add your response (or parts of it) to the post it's about?
Creative Commons Licence


From Dil Green, 2021-02-14, selected, with replies

Institutions: your description of Wikipedia as an institution seems to leave a slightly bad taste in your mouth. From a social anarchism pov, institutions are how society works. The framework of any persistent humn activity is always an institution. What matters is what it does, how it works, how it responds to change, how it dies etc etc. (of course, from an Illich perspective, institutions are straitjackets - words!. Fun note: I originated the wikipedia page on Ivan Illich - of course all of my words have now gone)

I am indeed thinking of institutions as necessary, though belonging to tradition and custom, therefore 3rd order.

'there is an algorithm' - hmm - in a world where fedwiki becomes something, this algorithm will be king - it is analogous, of course, to the ones FB and twitter use, and subject to all of the systemic realities of such mechanics. Its very existence is an admission of a mistake having been made, imho.

I agree that it is problematic. A possible solution in my mind is to systematically allow people to choose where links from their pages go to, though I am not clear exactly how to implement that.

'system helps you trace the original author of any passage' I have never believed in this. Having done a fair bit of collaborative doc writing, the fact that I put a 'the' into the paragraph, Bea put 'pangolin' in and Dee changed the spelling of wether to whether is going to make this meaningless pretty quickly (the history of my contributions to the Illich page referred to above turns out to be apposite, not simply an aside!).

Well, the information is there in the Wikipedia page history, it's just that it is not so easy to trace. I don't see any reason in principle why provenance should not be traced right down to the word and spelling level. Whether you want it is, of course, a different question.

'it is not yet well adapted to collectivity' the fact that this is still being promised, four years after i was introduced to fedwiki and told that it was a major concern, gives me little faith that it ever will be.

I would say, it is definitely not going to be collective, because that is not what its proponents want. A collective version would need either to be a pretty radical fork, or more likely a new implementation, as you suggest here …

Would we really need to start with Fedwiki? Lots of technical debt / path dependency in that community. One of the wonderful things about Ward's code is that he builds tools out of small, modules, each of which is beautifully crafted to support possibilities, then stitches them together. I think this means that it would be easy to fork modules that provided functionality in support of something which began with the 'group' paradigm.

Somewhere in the fedwiki riot/element channel there are links into a wonderful correspondence about 'choral' authoring of university level textbooks that you might be interested by.

Sounds interesting, I'd like to know more!

I'd be happy to collaborate on a doc which explored the 'what'. I would recommend hackmd.io as a free-form wiki-capable collaborative authoring environment for such work - it's almost capable of being a platform for what you are asking for - save that the governance would have to mostly human-mediated. This seems appropriate, though. Dog-fooding without needing code.

Anyone else?