This brings up a whole conversation around wikis and their usefulness. For the private knowledge base, I’ve always found TiddlyWiki to be an “ok” answer to some jot/note taking. With the dawn of Obsidian or bear.app (the one I use), something like TiddlyWiki doesn’t seem like something I’d choose.
For the days of USB sticks and traveling around to different computers, great, but in modern day TiddlyWiki just isn’t my cup of tea.
For “official” documentation, I always found it weird if not off-putting. Maybe I’m in a small camp, but through the years I’ve just never gotten to a place where it’s the first I think about.
I think for me TiddlyWiki is less interesting as a private knowledge base, and more interesting in terms of its architecture and the ways in which it is opinionated about hypertext and information more generally. When I first came into contact with it, I was put off, and found it to be ugly and kind of painful to use, but understanding its point of view helps a bit
In terms of architecture, the bit that’s interesting to me is that it’s programmable. The expectation is that you’ll write lots of small Tiddlers (“pages” or “posts” or “notes” – a tiddler is the smallest unit of information within TiddlyWiki), and compose them via linking and transclusion to build larger ones. So instead of writing a single tiddler called “to-do list” that lists particular todos, you’d do better to create individual tiddlers for each todo, giving them an appropriate tag, and then using the built in query functionality to build a list.
You can define macros and write Tiddlers that query other tiddlers in myriad ways using a declarative Prologish query language, and essentially make TiddlyWiki whatever you want or need it to be, kind of like Emacs – for better or for worse. In that way it’s essentially a programming language – TiddlyWiki itself is implemented in TiddlyWiki.
In terms of information, it embraces the ideas and idealism of the early pioneers of hypertext. This is evident in its terminology (the term transclusion was coined by Ted Nelson), and its support for backlinking, an oft cited deficiency of the HTML implementation of hypertext, among others
Another TiddlyWiki opinion is that your TiddlyWiki should be self-contained – without dependencies on external programs or websites that might change or break your TiddlyWiki. When you link to a Tiddler from within TiddlyWiki, you are linking to the object itself, so links and transclusions will follow the tiddler in the way you expect even if you rename it. This is also reflected in how TiddlyWiki lives in a single HTML file. This is in stark contrast to the brittle nature of HTML links, which don’t backlink, and reference an address rather than the object itself
But even with these “opinions” and best practices, TiddlyWiki is pretty dang flexible. Like the page I linked lists an ecommerce site, personal websites and information sites and documentation, and a game. It’s just HTML and JS under the hood
This is an interesting talk from the creators of Erlang and TiddlyWiki that goes into some more detail: https://www.youtube.com/watch?v=Uv1UfLPK7_Q
I agree, it really is a very flexible architecture
back in the day, before Wikipedia convinced everyone to see wikis as primarily for reference material, I also found wikis to be a really neat medium for conversation. there are very good reasons it doesn’t happen that way anymore (spam was a problem, and it was hard for newcomers to understand) but it was cool while it lasted.
I do use it. Image support is lacking but that problem also exists with anything Markdown-based.