(This was a reply to something on Medium)
From my point-of-view, anything with an API that I can map my own domain name to meets the bar of a resilient service.
D'Arcy Norman didn’t agree with me on that:
(hopefully that link still works. I mean, we don’t own Twitter, so that content could go away at any time)
I actually did export my Posterous content, and put it on my long term archive, which is self-hosted. Well, sort of. It’s not running on a server underneath my desk at home. It’s running on AWS. Is that self-hosting?
From my point of view,
I feel worse about the many many many people with SOMEBLOG.wordpress.com. You’re going to have to pay WP $13 per year (forever?) to redirect. If only they had mapped a domain name right away, their site would have been much more portable.
I recently added my own domain to Medium. I feel pretty good about my chances of exporting. And I feel great about doing the thing that we all worked on for so long — making it really easy to self-publish on the Internet. I’m writing more. I’m enjoying writing.
Moving the Goal Posts
Just today, Albert Wenger wrote about looking for a decentralized blog network:
But it is 2016 and if I am going to make the effort to move Continuations, it will be to something that I control entirely…ideally the platform I move to would include a decentralized network. Fred and I have been writing about protocol innovation and blog publishing seems to be one area that is perfect for such a system. In functionality my ideal system would be quite similar to Tumblr + Disqus but would use something like Onename as its identity system. Every post (and comment) would be signed with an identity and it would be possible to follow content based on identity independent of publishing location. For content storage it should be possible to plug in a decentralized file system via IPFS.
Absolutely! That’s what we should be working on next. Easy self-publishing? Check! Let’s move on to something that is even more resilient. A fully decentralized system.
We don’t even own our domain names — we rent them. And for the next 1 Billion, $10 per year is too much to pay to rent a domain name.
So here are my proposed rules for a decentralized, self-publishing system of the future:
-
Publish from a smartphone: someone with nothing but a smartphone should be able to publish.
This might be a full blown app (which could connect to the decentralized network directly, or even “host” from the phone) or just a mobile web page (which someone would need to run as a proxy / gateway into a decentralized system). - No server hosting: you don’t need to run a server under your desk or rent servers. Hit publish and your content is truly in the cloud — posted to a decentralized network.
-
No domain names: you don’t need to rent an identifier from anyone else to have the content be “yours”.
Remembering 0xA8C7372dC993d7510C9c45425807d463967cbb12 as my username isn’t going to be very usable, so there will be various user interface or even proxy systems that won’t be fully decentralized.
Consider this a first draft, and a way to start the discussion about where we should move the goal posts to.
I’m looking forward to a new wave of innovation around this. The discussion on Albert Wenger’s blog has a list of links that I’m checking out for experimenting with.
In the mean time, you can find me at tumblr.bmannconsulting.com and medium.bmannconsulting.com, my long term archive at bmannconsulting.com, and as bmann on the centralized corporate global messaging platform known as Twitter.
Further Reading
- Vinay Gupta’s Ethereum intro, Programmable Blockchains in Context
- IPFS, the InterPlanetary File System.
- More technical reading on trade-offs in decentralized systems, by Trent McConaghy, The DCS Triangle
- public document at doc.anagora.org/2016-08-01-self-hosting-doesnt-help-self-publish
- video call at meet.jit.si/2016-08-01-self-hosting-doesnt-help-self-publish