Starting a New Blogging System (create a title field under date to release)
I'm starting a new journal entry system with the 80-long hyphen as a blog-slicer token and place-holder for the Liquid template system. I'm using Jekyll, an open source static site generator, and a macro system to ensure reverse-chronological entries.
Creating a Journal Entry System with AI-Generated Content
By Michael Levin
Thursday, April 20, 2023
So starting a new journal entry is a little more yamlish than the previous system. It uses the 80-long hyphen as a blog-slicer token. It also uses it as a sort of place-holder for “New Page” and “Opening triple-hyphen” of the Liquid template system that Jekyll uses for front-matter. So it’s easy to imagine a blog like this as stack-combined individual pages that happen to have enough information to blend into the Jekyll post sub-folder blogging system.
----------------------------- 80 -->
date: Thu Apr 20, 2023
---
First, I’ll acknowledge that there are other static site generators out there, and some are even more Pythonic than Jekyll (such as Hyde and Hugo), but you’re looking for a deep sort of timelessness to your tools. You want to use APIs that plenty of other people like you also rely upon. Free and open source software first, please. And if not, proprietary is okay, so long as it can also host Linux containers like LXD for persistence.
No tech is perfect. And the only times you don’t step into what the mainstream is using is if it’s an excessively faddish or short-lived technology. Tech is full of the shiny new thing. Computers, the Internet, Mobile, Cloud, VR, and now AI. It’s helping me type this right now, hahaha! But it’s also a lot of noise. And it’s a lot of noise that’s not really helping you get your work
Haha, sometimes it gets it. Okay, I’ll get a little more work done. Thanks for the reminder. How does the pitch go? I like to keep a journal-file of my daily work. It’s a way to keep track of what I’m doing, and insofar as I can make the public. I only use Copilot on the journals I make public. And now with this super-charged help, I should be able to make it meta and use it to help me write the meta’s, right? Well, yeah mostly, but…
This becomes the basis for much work moving forward. It’s the ultimate fast-publishing systems. Get fast in vim. Have a macro system for starting a new journal entry. Make is so that it always works reverse-chronological, just like a blogging system. Make the system for parsing one page from the next, or one post from the next, or one journal-entry from the next, or what have you an 80 hyphens. That’s 80 minus. It’s sort of like homage to the old days of 80-column text editors. You’re always reminded of the optimum width to have usable text-editing when every byte was expensive.
The answer is 80-column, and it makes me happy to make that my parsing character for slicing the long-form journal.md, which you are usually ediiting, to a whole series of date-index.md files put into the posts subfolder of the Jekyll blog publishing system. It’s a data transform for 1-file to multi-files.
Now the multi-files follow a very specific format used by the Jekyll static site generator. If you want your content to be all linked-up like a blog for example, you need to have a date in the filename. You can add additional information after that, so I chose just a sequential index number. It doesn’t really matter except for in the git repo, because the “permalink” field in the front-matter will determine its URL on the site.
Want a way to start learning the vim/NeoVim text editors? Start with a daily journal. It will take the pressure off of you from learning to code. You can just clickity-clack away at the keyboard and get used to the motions. That’s what it’s all about in vim, getting it “into your fingers”.
Okay, so if you’re going to have a journaling system that you keep in one long text-file for life… well, don’t get ahead of yourself. Why would you do that? Okay, well the thinking goes like this. That’s the one natural order that’s going to exist anyway. You’re going to start making marks somewhere on something and the marks that you make are going to be in a natural order.
If you capture it in a journal, then you capture it in natural order. Whoa buddy, not so fast. Does that mean you’re writing always at the bottom of a text file like you’re always writing the last page in a book? Ugh, that’d be awful. No, you always want to be at the top of a file, pushing all the rest of the content down. We live in the digital age after all. Just put your new entries at the top of the file. So, every once in awhile, there’s got to be a time-travelling CHOP! That chop looks like:
CHOPPER = (80 * "-") + "\n"
And that looooong hyphen string is the token that gets used in a Python string.split() call to chop the file into a list of strings. Each string is yielded from a generator into one of two functions. The first creates the individual post pages for the Jekyll post subfolder, and a second rolls back in any new OpenAI content into the original journal file.
Once AI-generated content exists in the original journal.md file, it only gets updated from there, and in that way you can override the AI-written text.
Actually, I’m just in the process of testing that now.
It is my belief that this is one of the best ways to take advantage of an initial AI-generated text. You can use it as a starting point. In some cases it gets it right, and in other cases, you have to edit this or that. The danger then is losing that work. You don’t want such updates overwritten by the database.
Okay, I’m very pleased at how this is going. Give this one a title and let it publish. Then do some experimentation updating meta descriptions and such.