I would like to go a bit meta and tell you how I write this blog. This includes somewhat technical details and just general process. You can treat this is an overview of a life of a post in this blog - from an idea to being published. Thus, please let me welcome you to the backstage, but take your shoes off first.

Ideas

Every post in this blog starts with an idea. As you might have noticed, I don’t limit myself to any particular topic and this is also true for my ideas.

I collect ideas for posts non-stop. Whenever I get something vaguely resembling an idea for a post, I just take a note of it and add it to my list. The bar at this stage is extremely low. Basically, I just write it down (with enough info to be able to understand what I meant and occasionally possible ways to develop it) and forget. I use my “the worst that can happen” principle here (which I already described before). In the worst case, I just change my mind later and delete the idea. Thus, my goal here is to just collect something, evaluation will happen later.

This is my idea-collection phase (I call it “catching”). Overall, I like this process a lot, because it is hard to get an idea on purpose, but they do come often by themselves especially when walking or doing something unrelated. As a result, when I feel like writing a post, but don’t have any idea in mind, I can just look at my list and something immediately resonates with me.

This idea list has a lot of items at this stage. As I said, the bar is extremely low. I would say I have ~100 items there. Some are probably not actionable (the bar is unbelievably low here, did I say this already?). But there are 10-20 ones which I could just start writing about right now.

Obviously, this has some disadvantages. The list works like a stack. Thus, the oldest items just stay there and normally I don’t reach them at all. Normally I find something interesting already in the beginning of the list. With time I add more items and it gets even less likely to reach the oldest items. I could probably randomize the order, but this is not a problem for now. Just an observation.

I also publish some ideas as ideas in my “Future post ideas” list. As you might have noticed I don’t do this that often. Theoretically you could check out my ideas there and I would notice in my statistics that some of them are more popular than others and then write about the more popular ones sooner. In practice I haven’t got that much signal this way so far. Also publishing ideas feels cumbersome, because there are too many of them.

First minutes of post

Then at some point I feel like writing a post. If I don’t have anything in mind, I will go through my list and find something. Here I again use “The worst that can happen principle”. Basically, I just sit down and start writing whatever I like without rereading or correcting much. The idea is to get something out. Polishing is not a goal. In the worst case, I will reread it at some point and just delete the post.

Once I have the first reasonable draft, I put it into “baking” stage. This should have been named more like “aging” (like wine), but I like “baking”. I just leave the draft for at least a week without looking at it at all. The point is to completely forget about it. This is surprisingly easy.

Publishing

Choosing

I usually publish 2 posts per week these days - one on Tuesdays and one during the weekend. This is completely arbitrary. I definitely started with publishing randomly, but having some system helps to develop a habit.

Let’s say it is Tuesday and I need to publish a post. I just go into my folder with drafts and check whatever I have there. During good times I have 5-6 drafts at the same time (i.e. posts for the next 3 weeks already baking). During average times only 1-2 drafts (i.e. enough only for this week). Nowadays I am in the later stage. I should try to get to the former.

There are multiple benefits with having 5-6 drafts at any point in time:

  • posts are being baked for much longer (~3 weeks)
  • if something unexpected happens and I won’t have time to write new stuff, I will have enough posts in my reserve to get by for some time. On average publishing a draft is much faster than writing a draft. You can treat this as my emergency post fund.
  • when choosing what to post, it is easier to find something resonating with me when I choose out of 6 posts (instead of 1 or 2).

Post-baking processing

Once I selected something, I just go through it and fix whatever I don’t like. Very often the structure is weird or there are a bunch of typos or weird language constructs. Funny enough - this is impossible to see just after writing a draft and this almost screams at me when I reread the post after “baking” it. If the post is too bad or I find a major flow (e.g. false assumption), it can get deleted at this stage or postponed for additional improvements.

At the end of this stage, the post gets published! Hooray!

Post-publish

These days I also tweet about almost each published post in my twitter account. This is a good exercise for my mind to clearly identify the gist of my post. If you are into Twitter, this might be a good way to learn about the new posts. There is also an RSS feed, but I’ve heard that Google Reader is dead, so I don’t know how useful it is.

Occasionally I also publish on Reddit (my account). I do this only for posts which are relevant to the reddits I am aware of. There people can comment and I definitely enjoy this. Very often they propose corrections (e.g. I was thinking that capital gain tax with solidarity tax in Germany was 26.25% (i.e. 25% * 105%) and it was pointed out that it is 26.375% (i.e. 25% * 105.5%)). In such cases I do add corrections to my posts. Sometimes this involves redoing all the calculations and graphs (a lot of fun - irony^100).

Technical details

I use Jekyll framework to run my blog. I already touched the benefits of using Jekyll in an older post. Since I am into programming, Jekyll’s approach fits me well. My blog has “source code”, the same way as programs in programming. I also need to “compile” it to produce the actual final website. As a result, I can automate or simplify a bunch of changes. Theoretically, I could completely change the design and then recompile and all the pages would use the new design. Similarly with my “About” page - yes, it took me 40 minutes to set it up, but now when I add a new post, I just mention its category and everything else happens completely automatically.

As you might have already guessed, I also use git for version control of the blog. Theoretically this is not necessary for blog posts (especially since there is only me writing them), but as I said with Jekyll there is some programming involved, so I definitely prefer having git versions. This way if something goes wrong, I could easily go to a previous version.

I also love this feeling of seeing how something great and sizable is made step by step, a little bit at a time. Occasionally I look at my git log or blame and enjoy this feeling. This is also a good source of raw data for statistics about me writing this blog (e.g. how much and when).

The last section (normally this would be conclusions, but there is nothing to conclude in this post)

I hoped you enjoyed me giving you a peek into how this blog is getting created. That is a brief look behind the curtains. Please don’t forget your shoes. I definitely plan to continue with similar posts in the future (e.g. “Why I write this blog” and “Why I believe that everyone should write a blog”).

If you have any feedback, questions or suggestions, you are welcome to contact me with whatever channel you prefer (twitter, reddit, email - merelycurious protonmail.com). If you are unsure whether you should write to me, please use “The worst that can happen” principle. These days I don’t get that much feedback (I am going forward pretty blindly with the blog, mostly following my curiosity), so I will (with high probability) enjoy your feedback.

Happy going meta in your lives, folks!