View on GitHub

Quorten Blog 1

First blog for all Quorten's blog-like writings

What I learned from my switch to Jekyll blogging from my old fashioned text files process. The switch has been particularly interesting, so I’ll name a few items in a bulleted list form.

  • Switching directly to writing Jekyll posts is untenable if you are going to be writing a large number of discrete posts. For writing a large number of discrete posts, it is far better to write them all in one big text file, then use a script to split them up into the separate files required by Jekyll.

  • I wrote my simple script that takes a large text file and splits it up into individual article files with the assumption that I might primarily use it for migrating old data. However, I ended up using it quite frequently for writing my latest content too.

  • Writing the front matter on blog posts deserves special treatment. Although it is easy to write in the date, possibly via automatic timestamping, the other fields are quite difficult to write until you have finished writing your blog post. Also, the categories and tags fields deserve special thought that might not be avaiable even immediately after writing your blog post: categories and tags are best used when reviewing several blog posts all at once, when they have all already been written.

    That being said, most of the front matter fields are best left to being delayed until after you have finished writing your article. Don’t try to fill in all the fields right at the beginning of writing your article.

  • On the other hand, I still haven’t fully figured out what I want to do with categories versus tags. Ultimately, my primary use is more that of tags than categories. For now, both categories and tags are identical on most articles.

  • Indeed, one thing I have learned from practice in regard to categories and tags is that it often times makes sense to keep re-categorizing you articles. Often times you start out with a small number of articles that doesn’t require much categorization, but as you grow, what was previously just a single category, you want to split up into more specialized subject categories.

  • It’s important to have a catch-all “misc” category that all uncategorized articles fall into. This makes it easier to look back to find which articles should be reviewed for insertion into a more specific category.

  • One thing I like about Jekyll is that you have to create the categories and tags manually. It turns out this helps assure that you are reusing as many existing categories as possible to bin things together into.

  • Writing blog posts in Markdown is great!

  • Looking back on the text content once it is published in modern blog form, it sure looks and feels “smaller” than it did when I was using one big text file. Not to mention, the categories and tags organization mechanism works great! Well, at least when you use it properly.

  • Taking a large number of existing articles and assigning them titles, categories, and tags is easier than I thought. The main point of difficulty is when no dates have been recorded: often times it is possible to guess from surrounding entries, but this can be tough too.

  • One bad habit that I still need to practice getting rid of is writing articles that are too long. Specifically, when an article’s contents contain multiple different search subjects in a single article body. For the time being, I need to carefully review my articles and split apart single articles that should be listed as multiple articles. The big benefit of this is that it improves searching for information, by virtue of being able to use categories and tags generously.

  • Especially when operating in the field of technology, timestamps work quite well to determine what era of technology a particular article pertains to. Although obsolete technology is rarely useful for modern applications, it is always helpful to have accurate historic documentation when working with legacy systems.

    Again, I reiterate, because this is important! Also, as it turns out, humans are quite bad at perceiving time, so recorded timestamps are quite useful in making decisions.

  • GitHub Pages runs short on a key formatting feature that I must have for my blog: auto-generate hyperlinks from URL text.

  • Also, reviewing my blog, it would be really nice if when I cited a URL, I could have an auto-generated brief inserted too. For the time being, carrying over from my traditional notes process, it is at my discretion to add a short text description before the link if I feel such would be worthwhile.

  • Sometimes I forget to add the “more” divider tag to denote the end of the article brief view.