Leon Mika

Follow @lmika on Micro.blog.

Revisiting the decision to build a CMS

It’s been almost a month since I wrote about my decision to write a CMS for a blog that I was planning. I figured it might be time for an update.

In short, and for the second time this year, I’ve come to the conclusion that maintaining a CMS is not a good use of my time. The largest issue was the amount of effort that would have been needed in order to work on the things that don’t relate to content, such as styling. I’m not a web designer, so building the style from scratch would have taken a fair amount of time, which would have eaten into the amount of time I would have spent actually writing content. A close second was the need to add additional features to the CMS that were missing, like the ability to add extra pages, and a RSS feed. If I were to do this properly without taking any shortcuts, this too would have resulted in less time spent on content.

The final issue is that the solutions that I was trying to optimise for turned out not to be as big a deal as I first thought, such as the “theoretical ability to blog from anywhere”. I’ve tried using the CMS on the iPad a few times, and although it worked, the writing experience was far from ideal, and making it so would have meant more effort put into the CMS. In addition to this, I’ve discovered that I prefer working on the blog using my desktop, as I’ll more likely be in the state of mind to do so. Since my desktop already has tools like Git, I already had what I needed to theoretically blog from anywhere, and it was no longer necessary to recreate this requirement within the CMS itself.

So I’ve switched to a statically generated Hugo site served using GitHub Pages. I’m writing blog posts using Nova, which is actually a pretty decent tool for writing prose. To deploy, I simply generate the site using Hugo’s command line tool, and commit it all to Git. GitHub Pages does the rest.

After working this way for a week and a half, it turns out that I actually prefer the simplicity of this approach. At this stage both the CMS, and the blog it would have powered, has been in the works for about a month. The result is zero published posts, and it will probably not be launched at all1. Using the new approach, the new blog is public right now and I have already written 5 posts on it within the last 11 days, which I consider a good start.

We’ll see how far I’ll go with this new approach before I consider a custom CMS again, but I think it’s one that I’ve managed to get some traction now, particularly since it actually resulted in something. I think it’s also an approach I will adopt for new blogs going forward.


  1. This is not wholly because of the CMS: the focus of the blog would have been a bit narrower than I’d would have liked; but the custom CMS did has some bearing over this decision. [return]