Maru development is documented on an active mailing list.[1] Ohm development is being coordinated through GitHub. I'd personally like to take the extant code from OMeta/JS and the JS implementation of Nile & Gezira, and modernize them.
Recently I've been wondering if there's enough interest for a Discord server or something. (In the spirit of STEPS, it'd be ideal to make a new collaborative thing that's really different than static text/audio/video on the web, but gotta start somewhere. :) ) Unfortunately, I have had other, higher-priority projects at the moment, so I have taken no initiative to try to build a community.
I will also say that in my opinion, it's not clear to many of the people who made this stuff how special it is. The only exception to that is Bret Victor, who actually is not well-understood, but even the banana pudding versions of his ideas are typically much better than the industry's.
> I'm an outsider and also never got Frank to work. I was waiting for the Nile/Gezira thesis to get a high level (but hopefully also some detailed) descriptions) of how they handled graphics. I vaguely remember getting parts of idst working but for each of these projects, there were always multiple versions lying around. Sometimes in odd places.
I've never gotten Frank to work, and I abandoned my attempts. I've seen it run, though. The name was fully truthful: it really is Frankenstein's monster.
I did get Nile + Gezira to work (albeit in a very crude way by printing numbers to the console rather than hooking it up to a frame buffer). That's how I met Dan. I don't want to betray any confidences with him, but there is ongoing work with Nile.
Here's Dan himself presenting a related language at Oracle Open World in a demo (around 25 mins in).[2] (Full disclosure: I worked on the demo.)
If it were me getting started, I would take a look at the JavaScript implementation of Nile in Dan's Nile repo on GitHub. It should more or less work out of the box, and there's an HTML file containing a fairly full subset of Gezira. The only problem is that the JS style is way out of date, and so it does some things that are heavily frowned upon today. It may not work with tools like Webpack.
The Maru-based Nile is trickier to get working, but it does work. The issue with Ian's Maru is that it's quite hard to reason about and lacks clear debugging tools. I've gotten both up and running. I seem to remember the Boehm GC was pivotal in getting Maru to bootstrap and then run Nile.
> I read Alex Warth's thesis and it's well written, in a way that makes it very easy to understand. So, of course, I had to implement my own OMeta variant [2].
Pymetaterp is cool! I agree: Warth's work on OMeta was impressive. In some ways, Ohm feels inferior to me, though they're both good tools with lots of potential.
OMeta is the one tool from STEPS that is basically simple to understand and use without having to do a bunch of code archaeology.
> Also, the VPRI website itself says it's shutting down (presumably folks moved to HARC at that time?).
VPRI closed because STEPS ended and because Alan had to retire at some point. HARC and/or CDG Labs continued the work, but then closed as well. (I don't know all of the details, but someone here suggested SAP withdrew funding. That would track with what I do know.)
Today, Ian is teaching in Japan, Dan is at Vianai, Alex is at (IIRC) Google, Yoshiki is at Croquet, Bret Victor is doing Dynamicland, Vi Hart is at Microsoft Research and then Alan is retired. There were quite a few others I'm missing, and they are all doing interesting things as well.
[1] https://groups.google.com/g/maru-dev
[2] https://www.oracle.com/openworld/on-demand.html?bcid=6092429...