You are viewing archived messages.
Go here to search the history.

Konrad Hinsen 2025-03-24 06:10:46

My main recent FoC-related project is a framework for creating explorable, explainable, Web-publishable, composable, and convivial hypertext systems that integrate code. Combining aspects of explorable explanations, literate programming, and computational notebooks. The two principal intended use cases are (1) programs written for people to read, and only incidentally for machines to execute, and (2) scientific publications using computation.

A first demo is now available for browsing: hyperdoc.khinsen.net

This is what I will present at the upcoming meetup on Wednesday.

Maikel van de Lisdonk 2025-03-28 10:43:22

Last FoC meetup I showed the Open Canvas Interchange Format we're working at with the Open Canvas Working Group. We want to create a new spec to help (infinite) canvas apps enable interchanging their canvases with each other. More info can be found here canvasprotocol.org and this points to the latest version of the spec on github : spec.canvasprotocol.org .. and in case you want to help developing this: we meet every two weeks and here's the link to our next meetup upcoming tuesday : lu.ma/iv4q484x

📝 spec/spec/v0.4/spec.md at main · ocwg/spec

Open Canvas Specification. Contribute to ocwg/spec development by creating an account on GitHub.

Tom Larkworthy 2025-03-28 10:51:37

That was a great talk and very inspiring. The tight linking of apps and their canvas representation puts me off investing in layout algorithms because they seem so complex and limited in scope to the application they were develop for, which often does not justify the cost of developing them. My mind has been racing since about meta-layout algorithms that layout any diagram by any tool which makes more economic sense to me.

Now I wonder if you can take an ocwg diagram, convert to graphviz, let graphviz do the layout and then export out again 🤔

Tom Larkworthy 2025-03-28 10:52:30

Definitely want to play with this representation more.

Kartik Agaram 2025-03-29 19:48:43
Konrad Hinsen 2025-03-30 06:47:50

Nice!

Konrad Hinsen 2025-03-30 06:48:31

You don't provide much of a motivation for this, in terms of use cases, so I suppose the motivation exploring minimalistic hypertext design, right?

Konrad Hinsen 2025-03-30 06:50:50

The tacit design decisions are: (1) Hypertext is a set of pages in the same file system and (2) Lua as a dependency of the hypertext is fine. The second one is more fundamental here, because you could probably add over-the-wire links easily and compactly if your platform provides networking for you.

Kartik Agaram 2025-03-30 06:52:20

Yeah, I mentioned one approach that led me down this path in a recent devlog post: making codebases easier to explore by using a tiny markup implementation and also integrating some graphics. But I'm not entirely sure where I want to go myself. So this is all very exploratory.

Kartik Agaram 2025-03-30 06:53:18

I think the networking is easy to add. Depends on the use case. For my use cases so far the help browser is self-contained and doesn't need to download any resources over the network. But that will change.

Kartik Agaram 2025-03-30 06:56:18

I should clarify "help browser": If you try out the download at the bottom of OP above, it's a help browser that tries to document some bare essentials for programming on top of Lua Carousel, LÖVE and Lua.

Kartik Agaram 2025-03-30 06:57:09

So the hypertext browser is 600 lines of code, but there's also 700 lines of hypertext you can bounce around in.

Konrad Hinsen 2025-03-30 07:35:45

So the current/first use case is software documentation. That's a good one I'd say, given that I am doing that as well in my current project :-)