Decentralized

As I did twice before, I assembled many of my blog posts of several years into a curated summary which is now available from my Contents page. I called it “Decentralized Knowledge” although this is definitely not a correct title — if you have an alternative idea, please tell me.

There is certainly much more to be said about decentralization that I have not covered, in particular, the question of servers, addresses and hosting infrastructure. If you look at a web address there are many parts where you can have decentral instances instead of central ones. Compare

(1) http://central-provider.top/myown/page.html
(2) http://myown.central-provider.top/page.html
(3) https://central-platform.top/myown/post.html
(4) https://myown.central-platform.top/post.html
(5) https://myowndomain.top/post.html

My first homepages were of type (1) with the university and then with the formerly state-owned telco when there were no alternatives available, and I still have one of type (2) there. My first blogs were of type (1) and then type (2) at the university. Yes, corporate vandalism killed some addresses. This blog is of type (4) with wordpress.com, and I think this feels much more decentralized than sites on medium.com or hypotheses.org of type (3). Currently, I also have two of type (5): one on Reclaim Hosting and one with a German provider, with a self-hosted wordpress.org — but it’s still a bit centralized because I use Jetpack from wordpress.com, for services like trackback. Furthermore, the web server is shared and the database is on a central server — but the ‘server’ is probably on a cloud service which is probably distributed but maybe belongs to a monopoly kraken service.

(The little “s” in https is a hidden place for another hierarchical / central element: the security certificate must be signed by an authority that is recognized by the big browsers — and I still remember what trouble the German edu community had to get theirs acknowledged by Firefox. Not to mention our wires on the lowest level, which almost all meet at a big CIX center in Frankfurt, and the fat pipes across the Ocean.)

For me, the level of the site is crucial: if it is me who decides what to post, then it is decentralized.

IMHO, the bigger challenge is how all these addresses are communicated. Since the academic and public libraries have elegantly ‘delegated’ their task of resource cataloging to Google, we have now a monopoly. Similarly, for email and texting among friends and wider family, people’s addresses are no longer registered in telco’s directories, but another monopoly has built on that gap.

Such directories and registries cannot be decentral. But they are an essential part of the infrastructure, and hence should be publicly owned.

Again, the biggest problem is not on the server side but on the user side. And while the Google Reader gave access to decentralized blog resources, the Reader user interface was a centralized web application that was easy to kill, such that the promising many-to-many RSS topology is now almost dead, as well, or depending on broadcasting any new posts’ addresses via the centralized Twitter.