--- layout: post status: publish published: true title: Re-Decentralizing the Web wordpress_id: 2344 wordpress_url: https://www.martineve.com/?p=2344 date: !binary |- MjAxMi0wOS0wMiAxMzo0Njo0NyArMDIwMA== date_gmt: !binary |- MjAxMi0wOS0wMiAxMzo0Njo0NyArMDIwMA== categories: - Technology tags: - Technology - web - Hardt - Negri comments: [] ---

I have just seen, via Rohan Maitzen on Twitter, a useful page of suggestions for the "first day of term", teaching-wise. This led me to re-think a few of the ways in which I think the web should work.

I have been advocating, for quite a while now, the idea of re-decentralizing the web. This statement requires a bit of unpacking. As Michael Hardt and Antonio Negri flag up in Empire, the Internet "has no center and almost any portion can operate as an autonomous whole. [...] This democratic model is what Deleuze and Guattarri call a rhizome, a nonhierarchical and noncentered network structure". (Hardt, Michael, and Antonio Negri, Empire (Cambridge, Mass.: Harvard University Press, 2000), p. 299) Hardt and Negri, slightly problematically, use "internet" and "web" interchangeably in this portion of their text. The web, if defined as some form of hypertext conveyed over TCP/IP or other internet protocol and facilitated by DNS cannot really be said to be decentralized.

Firstly, the web, as defined above, and as used in most day-to-day operations by most users, requires DNS. The Domain Name System is a hierarchical system, not a decentralized one. Although delegation is performed to lower sub-servers, there are 13 root servers (albeit these are, again distributed via anycast) that form the centre. The web is not decentralized. Secondly, discovery mechanisms form centres of the web. Because the hypertext over TCP/IP portion that I described is decentralized, ways are needed to find content. More and more, these locations (Google, Bing etc.) are becoming the centres of the web, without which it could not function.

These are not, though, the primary sites of decentralization to which I am referring. I'm thinking instead more about content resilience, rather than overall network resilience. The link that Rohan posted is useful because it has solicited crowdsourced information and presented it in one thread. This centralization brings convenience. However, it lacks resilience. One single authority is in command of that space and could remove the information, or the entire site. Furthermore, it could even be the case that an event beyond the author's control brings the page down. All the comments are gone with it. Web 2.0 commenting features come with this danger.

The approach I'm advocating (that actually came about because I just don't like being forced to read other people's opinions, particularly on my own space) is quite simple and has, of course, already existed for a while. Don't comment. Use your own space and send a trackback (a notification that a page has been linked to) so that the threads emerge with links to other spaces. This functionality could be improved by some manner of script that pulls in the content, but that might be a bit tricky (identifying content, security etc.). In short: this is a better approach if you value resilience and autonomy over convenience... and you do, right?