To Compete with Medium

Dave Winer is encouraging bloggers (or really anyone with something to say) to post anywhere but Medium. He says that Medium is becoming a “consensus platform” for posting longform writing on the web, especially for people who don’t have a regular place to post. In doing that, Medium becomes a single point of failure, much like Twitter is for real-time short posts, or that Google Reader was for RSS. That means that Medium becomes increasingly capable of unilaterally changing how writing on the web works, for whatever purposes it desires. Medium could decide what you write, how it looks, who sees it, and whether or not you can take it elsewhere. And if Medium shuts down, you could lose everything you wrote.

Winer says that the reason people don’t just set up their own blog (even if they won’t write regularly) is because it feels wasteful to set up something and then not use it. This holds people back, even though a pure text blog takes up negligible space and bandwidth compared to videos or images. While he’s right about the minuscule size requirements of plain text, I think there’s more to users’ reluctance of setting up their own blog. There is a cognitive cost and mental overhead to setting up your own blog that Medium side-steps. To set up an blog on WordPress or Tumblr, you need to create a user account by providing a username, email address and password. Then you need to create the actual blog, by picking a domain name and title (and optionally, a theme). And then you can start to write.

Medium, on the other hand, lets you sign in via Twitter, automatically selecting your username and other account details (which you can change). After that you can just start writing. To be  fair, you are asked to follow other users and tags, but you can just click a button and move on. That’s exactly what I did before writing this post. There are options to use Facebook and email to sign up as well, but I’m assuming they’re equally streamlined. To break free from Medium’s hold on casual writing on the web, a competing service would have to be just as streamlined and painless.

So how would one go about competing with Medium? First you need to reuse identity from some existing social network or identity provider. Second, writing and publishing a post would have to be super-simple. Finally, to address Winer’s concerns, the competing service should come from an entity whose main business isn’t written content, but somehow naturally falls out of (or can be built atop) the core service. Luckily, there is already a service that can do this: GitHub.

GitHub is a popular code-sharing and hosting service that is very popular with programmers (and increasingly, with non-programmers). By default, GitHub hosts repositories of code, but they have an adjacent service called GitHub Pages that hosts simple websites. As a GitHub user, you can create a specially named repository and any HTML pages in that repository are served as username.github.io. Anyone with a GitHub account (which these days, is pretty much anyone who writes code) can post writing to their own repository and have it be served as a webpage from GitHub. Now, this only completes one part of the puzzle, since there’s no Medium-like interface to actually write your posts. You would have to write your posts using a text editor and push them to your GitHub pages repo. However, such an interface could be created by anyone, not necessarily by GitHub. They would just need your GitHub credentials, temporarily, to post your writing from the editor to the repository.

In conclusion: part of Medium’s attractiveness comes from having a streamlined path to posting irregular writing on the Web, helping to make it a large and powerful platform for web publishing. GitHub Pages provides part of the puzzle to create a neutral competitor that offers many of the same benefits. All that is needed is a writing interface that uses GitHub pages as a backend.

I haven’t talked about the social media and promotional features of Medium. I’m not sure how to replicate them in the same fashion. My goal with this post was to propose an alternative to the publish-and-forget style that Medium allows, and I think GitHub Pages is a step in that direction. Since Winer published his post, Medium has posted a response that addresses many of his concerns. The takeaway from the response seems to be that if you’re afraid of Medium having too much control over your content, post to both your own blog and to Medium.

Advertisements

Multiplicity

Over the last weekend I played around with an interesting service called Editorially. It’s essentially a stripped down, online text editor, with support for Markdown formatting. However, it’s most attractive feature is its support for multiple versions, collaboration and editing. It’s an interesting project and it just added WordPress and Dropbox export (I wrote the first draft of this post in it and then exported to WordPress). Like many such services, I’d rather use a text editor and git to get the same effect. However, more than the service itself what interests me is something I read on their blog post announcing their export features:

On the web today, a single article may be published on the writer’s personal blog, collected in an ebook, syndicated on several magazine or news sites, and reblogged across different platforms and communities.

This notion of having the same piece of work shared in multiple places is not new, but is becoming increasingly popular, especially with the rise of of group blogs (often with guest authors), online curation, and services like Medium. Craig Mod, whom I find to be one of the most insightful writers on the intersection of technology and publishing, started one of his recent pieces with this not-quite-disclaimer:

This was originally published in Hiut Denim’s yearbook. I’ve republished it here, over on Hi and over at Medium because, well, the beauty of the web is multiplicity. More on that later.

Multiplicity. I like that word. And yes, there is something to admire in just how easy it is to copy and share on this, the modern Web. But is it a thing of beauty? I’m much less certain than Mr. Mod, especially since this form of multiplicity is heavily dependent on third-party, often proprietary services with motives that are unclear at best and
questionable at worst.

What Mr. Mod dubs “multiplicity” and call beautiful can be explained in older, cruder terms: copy-and-paste. In many ways, the current web does us a disservice — we have been taught to accept and we settled for multiplicity when what we really wanted was transclusion, first described by an early pioneer of applied computation — Ted Nelson.

Whereas multiplicity on the web takes the form of copy-and-paste, transclusion would take the form of reference. Instead of taking the literal text (and perhaps the styling and images) of a document and replicating it for each copy, we have a single canonical copy of a document (and by document I mean any information object) that can be referenced and transcluded from other places. For example, Mr. Mod could have published the original piece on his website and Hi and Medium would simply transclude it in their own versions. When someone visited the Hi or Medium pages, they would reach out and embed the original post’s content within their contexts.

Transclusion offers many advantage over copy-and-paste. For one, any changes to the original are automatically reflected wherever it is transcluded. Second, attribution becomes much easier. Instead of carefully maintaining references to where you found a particular piece of information or text, the transclusion machinery can manage it for you. In fact, such a system needs to keep proper source information to work properly. Transclusion also makes the job easier for automated systems like search engines. Instead of coming across multiple versions of the same text in different places, a crawler would simply follow back the transclusion links and be able to index the original authoritative copy.

Copy-and-paste certainly has a place, even in a hypothetical transclusion-enabled Web. One major application is of course backup and archival, which would be impossible if there was only ever a single copy. That being said, personally I would rather have transclusion than not. For one thing it would make navigating the current morass of social media and publication startups easier.

Today, if I write something (say this blog post) and want to put it online, I have to decide where to put it. I could put it on my own website, hosted on my own server, accessible at my own domain name, where I retain full control. However, maybe I want the attention generated by publishing on a platform like Medium or Tumblr. Maybe I also want to post a link and excerpt to Facebook and Twitter and Google+. Once people read it, they might want to make posts about it on their blogs and reference it. It might get picked up discussed on random forums and message boards, discussion sites and Q&A sites.

To do that today requires a lot of manual intervention and thought. First I’d have to copy and paste the text into all the different services. Then I’d have to copy-and-paste a link into the various social media services. If I wanted to add an excerpt, I’d have to do more copy-pasting and editing. If there was discussion on other places, there’s no guarantee I’d find out about it. I’d have to keep a close eye on all the discussion sites, and hope that any individuals talking about it on their own sites send me an email (or some other kind of notification) about their posts.

In a world where transclusion is the default, things become much simpler. As we’ve already discussed, the various other publishing platforms would simply transclude the content of my post. Social media services would transclude particular paragraphs (or even particular sentences). Similarly, discussion sites and other people’s blogs would transclude the particular parts of the post they want to discuss. This has a secondary benefit: I can look up the transcluders and automatically be aware of who’s talking about my post (and what parts in particular). In summary, transclusion would make sharing and discussion on the web a whole lot easier, smoother and interactive.

Unfortunately, I don’t believe that we’ll achieve transclusion any time soon. In particular, I would say most publishing and social media services have incentives to prevent transclusion — they want a unique piece of your work. Deferring to a canonical copy elsewhere that others can transclude as well is the last thing they want. That being said, we can still dream, can’t we? Perhaps, with the continuing popularity of ebooks and DIY publishing we might even start having some limited forms of transclusion. And maybe, just maybe, people like Mr. Mod and services like Editorially will start pushing for a transclusion-capable world.