MYWORD.IO: Separating the Application from the Data


This post was originally published on my blog on February 20th, 2015.

I’m intrigued by, a new publishing tool created by Dave Winer. It’s simply a web page that pulls in some JSON via an url parameter – so you can host the content of a web page virtually anywhere (or at least anywhere that supports cross-origin resource sharing).

Phil Windley writes:

This is an example of what I’ve been calling a person cloud application architecture (PCAA). The key idea is to separate the application from the data and allow the data to be hosted anywhere the owner chooses. The advantage is that there’s no central server intermediating the interaction.

There are intriguing possibilites when you separate the app from the data. For one, anonymity. won’t break even if you leave out a value for authorname in your JSON. And if you host your file on dropbox (or as I have enabled with myword-json-converter, saving your JSON to, nobody knows where the data is really hosted. (Meaning, yes it’s on dropbox, but whose dropbox?)

There’s both a freedom and danger here. There is no way to confirm authorship unless the source of the JSON file is confirmed. For example, if the url parameter references a file on then there’s proof (unless someone hacked my site) that I uploaded the file there, so I can’t post an essay on and say it was written by Ev Williams and not be challenged. But if I reference a file hosted on dropbox or, there really is no way to validate authorship. Obviously anonymity confers freedom to share thoughts without retribution, but one can make it appear that your essay is penned by another person – which is bad for reputation management.

Another downside to a tool like this is search. I don’t see how Google can “read” the contents of the post, since all the content (except that which is in the “) is hosted remotely. (The text is not in the HTML of the web page.) If the contents of the web page are not crawlable, you can share an URL privately, and only those who know the URL will have access to the contents. But this is not a publishing platform you would want to use for building a web presence or getting traffic through search. can be great for sharing, but not for discovery.

Essentially is to JSON what SmallPicture Reader is to OPML. They’re just displays for content hosted elsewhere in the form of structured data. They’re convenient in that – provided you have the knowledge or tools to output an essay to JSON or an outline to OPML – you can share it with readers. You don’t need a complicated CMS. You just feed an URL to either of these services as a parameter (?url= or ?opmlurl=). You don’t need to use a script to generate static pages, or store your data in a MySQL table. Just upload the file somewhere to the cloud and have the app read your file.

Dave Winer has open-sourced, so you can install it on your own server. And you can host your JSON file wherever you like. This is empowering, because you now are completely in charge. You don’t need to host your content on Medium to publish a nice-looking web page. You don’t need to run a complicated piece of software like WordPress that is dependent on the LAMP stack. runs right in the browser. But I wonder at times about the implications of separating the application from the data.