2023-07-08 A summary of my journey to Oddmuse

Article describing an evolution in thought process that led to using Oddmuse for my blog and site.

[2023-07-03 Mon 23:07]

Tags:

As I review my notes and journal entries on OddmuseWiki, I noticed a clear change in thought pattern over time, and this is an attempt to quantify the same.

Towards the end of 2019, I managed to get my first oddmuse wiki setup using an Apache webserver. I was able to setup multiple wikis, with the idea that I would use oddmuse for my blog and consider separate wikis for different purposes.

In parallel, I was also experimenting with using Wordpress (again), and the indieweb plugins which enabled publishing at a single location and having that content federated to other platforms. Though the latter was extremely interesting and somewhat fulfilling - it felt like my blog was a bit too intensely connected to social media than I was comfortable with. In addition, it seemed to impose a certain 'expectation' of the kind of posts I would put out there.

While veteran bloggers may stand by wordpress purely in terms of features and the ease of managing a CMS - the typical issue with wordpress is also maintenance and bloat. That being said, a database based CMS seems like overkill for a personal website without multiple users.

With respect to Oddmuse - while it easy to get started with pages and interlinking pages - it was hard(er) figuring out a blog, and much harder in getting everything to look as nice as the Hugo blogs out there. Needless to say, I failed at finding a constant motivation to improve the web interface - it felt like a perpetually growing, endless list of things that could be done to make things 'look great'; while a voice whispered that 'content is always king'. Some great examples of simple, content focused design are the blogs of Paul Graham, Alex Schroeder and others like and All Things Distributed (the blog of Amazon's CTO, Werner Vogel, as well as Protesilaos Stavrou's and Sacha Chua's site.

My approach back in 2019 and continuing to a few years later - was to rigorously hunt for Hugo themes and try to find the one that satisfied most if not all the features that I wanted, or atleast one where this could be achieved with minimal modifications.

I found such choices often resulted in bloated sites including a lot of Javascript and other 'things' that I did not want to mess with unless needed. One of my criteria here was also that the site should render well in a text based browser.

It should be noted that I did also find great looking minimal Hugo themes which I tried to hack on during different periods of time. The results of this exploration-from-scratch were not significantly better than what was possible with oddmuse out of the box. Not that I was ever finished, and I did make some progress, but it was usually tedius and hard to keep up with - and most importantly, it took away the time I spent 'expressing' via writing.

The ox-hugo approach to publishing using Emacs is popular for many reasons, besides offering a streamlined workflow. From a single (or if desired, multiple) Org files - it is fairly easy to manage, filter, tag, categorise, edit and link between blog posts, as well as pages. One can integrate that org file into the agenda and use features like clocking as well, and it's worth noting that Hugo can render Org files directly.

Using ox-hugo worked rather well initially. But over time, as my note taking systems and perhaps thinking style evolved - from simple date trees in a giant Org file, to mutiple agenda files, to OrgJournal? (one file per day), and then OrgRoam? (one file per note) and finally EmacsWiki:HowmMode (staying with one file per note)-> my tastes also evolved to want a simple website, with as little of the ubiquitous ‘selling’ and bloat as possible. I also desired a certain flexibility to organise content and interlink them with ease.

Conforming with the recommended style of ox-hugo, my personal preference was also to have a single source of truth, a single Org mode file that would be the source of all the articles on mysite. This approach also means that one is determined to only edit the source, and republish any time the content is to be updated. i.e. the thought had to be precisely captured as a task - to add to the infinite to-do list.

However - way too often for comfort, I've found myself wandering somewhere away from a computer, and finding a typo in an old article or wanting to add a super quick note or caveat to a past post - and being unable to do so on the spot. Wikis, even with a simple web interface for edits are often more conducive that kind of editing. I disliked the impediment towards a 'quick evolution' of an article with static sites, and this was not so much an issue with an OddmuseWiki, though ofcourse, the web editor and the mobile experience still leaves much to be desired.

While using emacs : packages like oddmuse.el, oddmuse-curl and yaoddmuse.el provide the a simplistic pleasure in directly editing the wiki content in a buffer and being able to preview right within Emacs. Since there is no longer a single source of truth - an oddmuse user would have to diff between local notes and the oddmuse wiki page. This can be easily accomplished applying diffs using Emacs tools.

I have also repeatedly noted in my journals that the 'wiki' as such is designed to invite collaboration, which if it ever happens - should in theory (mostly) improve the content or atleast lead the way to interesting discussions.

What emerged rather strongly was that this had to be a space for me to express myself, rather than sell myself or sell any particular opinion.. From these perspectives though, both Hugo and Oddmuse qualify. However, I think Oddmuse still has a bit of an edge here since collating different kinds of content is a lot easier than it is with Hugo, since it is often based on a text search.

After a hiatus in actually publishing the things I write, somewhere around 2020 - my return to OddmuseWiki in the last few days has revealed a certain balanced comfort in using a wiki: Everything can be made better (in time or through collaboration) - but it has to exist first - and it has to be published for it to exist. In addition, I re-discovered that my so-called home grown Zettelkasten had a tonne of rants, notes and other expressions on the topic of Oddmuse. This article was cobbled together reading the above, just in the way the atomic notes on a topic are supposed to 'come together', but that's for a different post.

Beyond all these things - oddmuse powers the EmacsWiki, where I am currently one of the admins. Learning more about Oddmuse (and hopefully contributing at some point) - would have the added benefit of giving something back to the community.