Tercio de Melo Sousa

↵How I write

Continous Delivery is a software development practice that I summarize in two points:

1. Release when it's good enough, not when it's good.
2. Release new versions very frequently, at every minor improvement or change if it doesn't break your software.

That's how I write. This is not the first time that I try to keep a blog going, and besides my othe few attempts I had quite a few personal software projects that were abandoned in their infancy. The reason, I think, for their premature death was the lack of Continuous Delivery.

Release when it's good enough

I've come to understand that I if try to polish enough to achieve a great writing quality bar I will not publish many texts. In fact I will be discouraged to start a text by predicting the effort that it will take to make a publishable version of it, and then the overall quality of the project will be lower than if I had published lower quality texts. It's better to get anything done, than to do something perfectly.

What does it mean to be good enough, though? To me, if the text content is complete, it's good enough; no metter the shape. This blog's goal is to tell stories, and if a text is telling the whole story, it's good enough to be published, even if it's not telling this story as well as I would like it to.

I write my texts first in English. The only thing that I do between them becoming good enough and publishing them is translating them to Portuguese, which gives me a very early opportunity for improving things as I go. I publish it once the portuguese version is ready.

Release new versions very frequently

After publishing I spend up to an hour going through the text, improving, rewording, rearranging, pruning uneeded words, sentences and ideas, fixing grammatical errors, etc. At every such change I relesae a new version of the text.

That's how it should be. This is the internet, I'm not writing a book that needs to go through an editor and a publisher, that will be printed in thousands of paper copies and sold in bookstores. Fixing a published text is cheap and fast, there's simply no value in trying to publish a very high quality first version; specially given that I'm not being paid to do this.

At every version published the text gets better and the next reader will have a better reading experience.

A day after publishing I'll give another look on it. With a fresher mind I will see opportunities for improvements that I couldn't see before. Over the course of the following week I may revisit the text a couple of times to do some more polishing if I think it's needed.