Eight months ago, I decided to try Medium as the platform on which to post my essays. Over this time I have published a handful of posts in there—8, to be precise, which is… a very shy number—but the results have been quite satisfactory: the WYSIWYG composer is excellent, the analytics tools are simple but to the point, the looks are great, and the community is nice (though I haven’t been able to tap into it just yet). But where have things failed?

The failure

Regardless of all the good things, the experiment overall has failed. To summarize it:

Medium has failed as the place where I want to post my articles first though it remains a great secondary place for content redistribution and promotion.

Let’s go over the reasons.

Post classification

The first problem I have with Medium, which is a minor one, is that I see it as a place for high-quality, large-form essays—not the average blog post with the musings of the day. In other words: I see myself posting long opinion pieces (as I have done) and not something like the announcement of a new release of a personal project or incomplete thoughts on a topic. In turn, this means that I have restrained the kind of content I have posted there and thus not been able to completely replace my blog.

It’s completely true that this is my own point of view and I’d just need to let go of it and post more. But there is an implementation detail of Medium that bothers me enough to prevent me from doing this, and it is a simple one: posts cannot be classified. As a result, as soon as you start posting short replies to other people’s posts, your own profile gets cluttered with such replies-as-first-class-posts and your original, long-form content gets lost in the noise.

It would be great if Medium provided a mechanism to say “show these and those stories on my profile, but nothing else” so that you could carefully craft what you want displayed on your persona. Otherwise it is hard for readers to see what other good stuff from you they may consume. Maybe a publication is the answer to this, I don’t know, but I don’t really want to dig further because of the second problem.

Data liberation

The second problem, and this is a much more serious issue, is about data liberation. If you write first on Medium, your posts are stuck there unless you put significant effort to take them out.

But why is so? After all, it’s just text! Well, you see… Medium provides functionality to export your content, but the exported posts are unusable garbage: the exported HTML can be read on a browser and it renders similarly to what you could view originally in Medium, but that’s about it. The HTML is littered with unnecessary tags, and to prove that suffices to say that the exported HTML for a post takes double the space than the raw text. As a result, the exported copies are unusable for anything else should you ever want to reuse your text for other purposes or in other venues.

Importing the posts into this site

For these reasons, I have spent a significant amount of time “taking out” my posts from Medium and reformatting them to fit this site. The result are Markdown posts that are easier to deal with and are future-proof: the Markdown format is simple enough and readable enough that one can imagine the text being perfectly usable and parseable years down the road.

This was an excruciating process and I was lucky to only have 8 posts to export. I started by using html2text to convert the Medium export into Markdown files. Then I had to manually strip out Unicode characters and fix spacing issues in the text itself caused by their presence. Then, the worst part, I had to deal with images and embeds. And, lastly, I had to update this site to tie things together properly.

I do intend to keep posting to Medium but the process will be the other way around: first write the master copy in a format I control and then import it into Medium for promotion. Maybe this will prompt me to update the Markdown2Social tool to support direct publication to Medium.

For your reference, here is the list of posts taken out:

• 2015-10-23: Compilers in the (BSD) base system
• 2015-09-24: An open letter to online support staff
• 2015-09-20: “Your English is pretty good!”, they said
• 2015-09-07: My coding workflow
• 2015-06-19: How to commit a code hack and not perish along the way
• 2015-06-06: Get a handle on email subscriptions
• 2015-05-26: Six years at Google
• 2015-05-24: Hello, Medium!

What about The Julipedia?

Unfortunately, the complaints above do not apply to Medium alone.

The data liberation issue applies equally to Blogger, and thus to The Julipedia. It is true that Blogger is in a better position than Medium because one can compose posts directly in HTML (as I have been doing) so you can actually take out almost-exactly what you put in. Unfortunately, Blogger has a tendency to mess up the HTML you type if you happen to embed verbatim text or use the WYSIWYG editor by mistake. Therefore, I am also thinking of taking some of the posts out from there and putting them here… but I will not call The Julipedia dead until that happens.

To tie things together, I have done a few improvements to this site and made the blog section a first-class citizen. In particular, I have added excerpts to all the posts, improved the posts index, and added commenting support via Disqus. (What? I am complaining about Data liberation above and I still resorted to an external commenting system that I do not run? Yes, that’s a fair point… but I’m not ready to run a dynamic web site yet.)

Subscribe via RSS right now to not miss a beat. And if you happen to read this, would you provide the very first comments to this site via the comment box below?