Version 3

I decided to finally get my act together and update my website. Of course, me being me, I couldn't just update the events page and call it a day. No, I had to do a complete rebuild. It's not a redesign, because it's not just visible style tweaks, but a complete overhaul of how the site is built and deployed, as well as reconsidering and greatly expanding the purpose(s) it serves.

I miss writing online, I miss the feeling of processing my thoughts as I get them out of my head. And while I have been doing a pretty good job of writing in my pen-and-paper journal over the last few years, it's not the same as writing for a (potential) audience. Additionally, there is a difference between writing slowly by hand and quickly smashing out some words on a keyboard. I was also a little concerned that instead of writing longer, considered posts and essays, I was just occasionally writing brief tweets. Twitter still has some value (as long as I avoid falling into the doomscrolling trap), it's certainly not the only place I want my writing to be appearing.

It's taken me a while to actually get back to it. In the past, I've had a WordPress blog that I've used on and off for over a decade (yes, it feels weird to be counting time in decades now). But over time, WordPress changed, and the impedance mismatch between the way WordPress worked and the way I wanted to write continued to grow. WordPress is (perhaps understandably) a content management system, a way for groups to manage websites with lots of content and integrations. It's no longer really a tool for one person who just wants a quiet environment to write and think in. I might still recommend it to beginners who want to set up a good looking website or blog for little effort, but it's not for me anymore.

I briefly tried out Micro.blog and Write.as. They both have their positive points, but while WordPress was too much, both of them were ultimately too little. I needed a publishing platform that worked the way I thought, and that I could improve and alter when I came up against its limitations. In 2020, partially spurred by pandemic ennui, I started working on my own site builder. While I hope that one day I will bring it to fruition, right now I want to get back to actually writing, rather than working on my writing tools. At the same time, if I was going to move away from hosted services, I wanted to move ahead in a way that established a firm foundation for whatever direction I wanted to go in the future.

As I was thinking about all of this, I had the good fortune of coming across a number of websites and blogs by people who seemed like kindred souls. Their experiences, outlooks, and decisions informed by own, and helped me decide on a path forward. First off, there was Simon Collison reminiscing about how our websites used to be our playgrounds, and the benefits of maintaining a lightweight "stream" in adddition to a more weighty blog. He also pointed to Rob Weychert, whose thoughts on using his site as an archive of his online activities definitely mirrored what I have long wanted from my own site (in addition to providing some concrete technical details and ideas about color). And last, but certainly not least, James Larisch and Andre Garzia encouraged me to embrace chaos, and get started with writing and publishing, even if there were some rough edges.

So here I am, at version 3 of my website, embracing the chaos and incompleteness, working with something that's part playground and part archive. All of which leads to the question:

Why Version 3?

I'm counting version 1 as the hand-rolled HTML version of the site that started around 2014. Version 2 is what's lived here for the past few years, mostly static HTML pages, generated using the Pollen static site generator.

Writing HTML by hand is fun and all, but right now, I'd rather focus on actually writing. And while Pollen is a very powerful way of writing web-based documents, it is currently more suited for what are essentially books-as-websites: artifacts like Practical Typography and Beautiful Racket. One day I would love to help it reach it's potential as a tool for building regularly changing personal sites (blogs, wiki, personal knowledge gardens and the like), but today is not that day.

I wanted something that was essentially static, generating mostly HTML pages with some CSS, which could be served by something like Nginx and easily backed up. I wanted to be able to edit posts locally, on my machine, even when offline. Ideally, I would be able to use my customized Spacemacs setup to edit quickly and efficiently. But I also wanted to be able easily preview pages and posts before I made them public. I wanted to back up said posts with a version control tool like Git. I initially tried to use Blot.im, which has a dead simple deployment model (based on Dropbox, or Git, or Google Drive). However, Blot is really optimized for blogs, and I wanted something where I could have separate sections that were not blog-like (see the Notes section, including my evaluation of Blot).

Ultimately, I settled on using Hugo, a fast static site generator that does about 90% of what I want it to do. The site is mostly written in Markdown, but with custom templates and CSS to get things looking the way I want. I can have a number of different sections that work differently. Currently I have a journal that is essentially a blog (I just don't like the term "blog") and a notes section for standalone pages that works similar to a wiki, or personal digital garden would. I can write new entries in either section in Markdown, and Hugo will translate them to HTMl and update the indexes automatically. Hugo also has a nice live preview feature which lets me see the generated HTML every time I save changes. Everything stays local to my machine until I decide to push things to my server (for which I'm using good old rsync). Hugo's default folder structure also lets me mostly separate out the content, styling and generated site, and back them up as separate Git repos.

Like I said, Hugo supports about 90% of what I want it to do. The missing 10% includes things may grate on me over time, but that I can live with for now. Markdown has its limitations (which will make for its own post). The templating system is quite powerful, but takes some getting used to, and there are some limitations that feel arbitrary and awkward. While I can keep a (mostly) separate notebook and journal, things like tags and categories are shared across the site. This means that journal entries will currently show up in the notes index as it just shows everything by tag. In Hugo's defense, I haven't actually sat down to try to fix this issue. I've been keeping more detailed design notes, if you're interested.

I suspect I will continue tinkering with Hugo and the site as I continue to use it, but for now it seems to serve its purpose of mostly getting out of my way so that I can get back into a rhythm of writing.

What's next?

I've long wanted my website to be a way of reducing my dependence on third-party services, all of which will almost certainly become unavailable at some point. While I want to continue still using them (mostly for network effects), I want this site to at least be a mirror and archive for what I post elsewhere.

To that end, I've taken a couple of pages out of Rob Weychert's book. The site is somewhat future-proofed by redirecting visitors to a particular version of the site, i.e., going to https://basus.me/some/page/ takes you to https://v3.basus.me/some/page. Since the generated HTML is saved and archived independently of the source, there's a greater chance that I can keep versions around, links intact, or restore them easily if they get lost.

I want to gradually move things I've posted to other services here. Things I've put on Micro.blog and Write.as will probably come first. They are the most recent things I've written and there are few of them. The ByteBaker will move over gradually, but there are over 500 posts, so it's going to take a while. I have a decrepit old Tumblr, most of which is now dead links, but some of which is worth saving. Eventually, I would also like to archive and mirror some parts of my Twitter and Instagram feeds. But I'm debating as to whether those kinds of very short posts should live in the journal, or a in a separate stream section. There are also a few larger essays here and there that should be brushed off and put in their own section.

At some point, I will sit down and figure out the RSS situation. I'm still a big believer in being able to subscribe directly to a site, rather than getting updates via Twitter or a mailing list. There will probably be a couple of different feeds, one for each section, and a unified feed if you really want everything. It may be nice to share private drafts in some way, I haven't thought about that yet. I'm not super interested in analytics at the moment, but it might be useful to have some idea of how many people are visiting my site, and why, in a privacy-preserving way. I suspect that design tweaks will continue indefinitely.

Ok, now what?

For the near future, I would like to to a routine of posting 2-3 times a week. Over the past few weeks, as I've been working on the site, I've done a pretty good job of keeping and updating notes, and I hope to continue that. I suspect that some notes will eventually feed into blog posts (for example, I have more detailed notes on the v3 rebuild). Hopefully, every now and then I will have the time and space to write a longer essay. Ultimately, I would like this site to support not just my writing online, but a lifestyle of thinking and working in semi-public (in a "public intellectual" way and not a dystopian surveillance way). The site is one part of that, the easiest one to get up and running, but no spoilers!